Friday, July 31, 2015

Dota 2 Next Level Meta: Support Broodmother

I have been putting this one off for the longest of times because... I fucking hate playing Broodmother. Scratch that, it's not exactly true, because I like the hero conceptually - it's just that playing him in pubs is miserable. In fact, just as miserable as playing with her on your team usually. So, how does Brood play as a  support? My initial was "trash", but let's find out...

Hero analysis

Let me start by telling you why Brood is such a miserable pub hero: of all the heroes in Dota, she's the one that defines the game most just by being picked. She excels at conquering a lane and pretty keeping it as her own unless half the enemy team is constantly present trying to keep her off. So she's not just a spider, she's also a filthy rat - the rat of rats. There are two problems with the hero for pub play: 1) she's garbage in teamfights and 2) it's not enough for the Brood player to be good at Brood Dota, the rest of the team are also forced to play Brood Dota - and usually they don't do that very well. Brood is very annoying and powerful in small skirmishes where she has all the time in the world to just mess with people, but in teamfights with big ultimates she's a super squishy melee hero who can't very well function outside her webs. Recent patches made it easier for her to at least deploy more than one web in a fight but she's still really squishy.

The hero is pretty much entirely defined by her webs. Regardless of the role you play, webs give you the ability to move around effortlessly, creating your own paths, and regenerate health. On level 1 you get two webs, which is enough to cover the basic laning situation decently. As a support you can use this mobility and free regen to harass your enemies, until they get annoyed enough to bring sentries. Your harassment is pretty crap though because Brood's base damage is kinda low, and you still have to be in melee range. Once you level them up, you can also use webs to transition between two lanes and generally move around very efficiently. Her other signature skill is Spawn Spiderlings which does damage and spawns spiderlings (who'd guessed). This skill combined with her webs makes Brood a pain to play against, because those spiderlings can actually hurt you a bunch and then just vanish before you can kill them. Well, they are also a good way to feed gold to enemies if Brood isn't careful.

Since her nuke gives you additional units, it sounds like it might be suitable for jungling. However, the damage on level 1 is crap, and it only spawns one spiderling which you need to keep alive in order for it to spawn spiderites. I will test the jungling capabilities of this skill. Her third skill is probably her best early game harassment skill. The passive gives her attacks a slow and causes a miss chance for enemies, which makes trading more favorable for Brood. It's also a nice way to chase your prey. Unfortunately it's not that amazing as a slow, as the level 1 slow is only 10%. This is where the biggest problem of playing support Brood comes into play: all of her skills are terrible at low levels, so you want to have all of them leveled up as fast as possible - and leveling up fast is exactly what supports generally don't do. We'll talk about this a bit more later. Before that, we still have her ultimate to discuss.

Not that there's a lot to write home about. A damage bonus for your melee attacks is generally not very useful for supports, because as a support Brood there is no way to sustain yourself - even with that juicy lifesteal - in melee range for long. Because the other skills are so important to level up, I might even consider skipping her ult at level 6 and grab it later.


There is one early game item above all: Orb of Venom. Especially because you're a strong independent matriarch who don't need no boots. The movement speed bonus she gets in her webs is more than enough for laning and pretty much the entire game - indeed, once she has maxed webs, her movement speed is already very close to the maximum. Of course you are really fucked if should you ever leave your webs, but why would you? So, assuming we're laning, and don't need to buy wards (wishful thinking I know), it's safe to start with Orb of Venom and immediately go towards an actual item, skipping boots. Saving that 450 gold is huge for supports. What items to build though? Urn is probably not that great on Brood, because even as a support you will probably eventually pack your things and make a home in the enemy's safe lane and jungle. If you can get early game kills it might be a worthwhile purchase though. At least it gives you some HP.

Itemization is really highly dependent on what you're going to do... as support Brood. I have no real idea here folks. If you're going to use your nuke at all, Soul Ring is a necessity. Mek is probably wasted on you, because you won't have the mana to use it, and again, are you actually going to be in teamfights? Vlad is a more suitable item, because it's generally helpful whether you are alone or with your team. Eul is such a trendy item these days that I'd consider grabbing one for initiation and to do the one thing you can't: cancel TPs. If you get one, you definitely have no use for boots. Force Staff is another useful item for fights. Blink is also useful to escape ganks if you're splitpushing. You can always also go full greed and aim for Dagon, Orchid or Necrobook. In this case you might as well grab a Midas to get there, because the levels will be yummy as we discussed earlier. Orb of Venom into Midas might actually be a thing for support Brood. Make a stop for Soul Ring if you want to jungle.

Game plan

I see two routes to take here, and I'm not sure which is worse. Probably both. The first route is to play like a core Brood with a bad start. In other words, you start as a support, and once your farming core no longer needs you, you recover by starting to farm. Ideally your core gives you the lane and goes elsewhere. This was demonstrated by SyndereN in one episode of Dota Cinema's A-Z Dual Lane Challenge when they played Broodmother and Centaur on a dual lane. The other option is to try and play a ganking/utility Brood, where you use webs to move around the map and your slows to help in ganks. In this approach you also take parts in fights by throwing nukes and using any utility items you may have been able to grab. The problem with the first approach is that Broodmother is a shit late game hero, and can't really do anything if she's behind because then just one enemy support can easily fend you off or even kill you when you try to splitpush. The second approach is also bad because Broodmother is a weak hero without levels in general.

The start of the game depends on whether you can actually jungle or not. I will test this, but I'm not very optimistic. If you get lucky with the small camp and get the one with 5 creeps, you can get a small army of 1 spiderling and 5 spiderites, which may be enough to move on to a medium camp (the satyr one is ideal). You can also spawn more spiderites if you pull the safe lane, and hit your own creeps with your spiderling before they die. However, because spiderites can't spawn more spiderates, you will eventually need to use your nuke again. Having a few clarities is therefore required, but it might be hard to tank the neutrals with spiderites while you are trying to clarity your mana back up. You'd also need to start entirely without webs, which sounds a bit hazardous. Unless you take webs at level 1 and try some chokepoint shenanigans. Once you get something like level 3 and a Soul Ring jungling becomes quite effortless I'd think. In this case by level 5 you could have 3-2-0-0 skill build. Makes you a shit ganker, especially if you skip Orb of Venom to get a faster Soul Ring.

If laning, I think prioritizing Incapacitating Bite is better. In this case you should start with webs to give yourself an actual laning presence, and then put the next two levels into bite. This combined with your orb gives you a notable slow, which your carry can leverage to land a stun or something on the enemies, and then you can hopefully kill them. Your contribution to actually killing anyone at level 3 is less than that of a support Riki in terms of damage though, so don't get your hopes up. So you better be laning with a carry who has high killing potential if they get to hit the enemy a few times (e.g. Ursa). The slow from level 2 bite should be sufficient for further ganks, so beyond that consider leveling your webs up to get more mobility on the map. Then you can proceed to become a nuisance on two lanes instead of one. This approach is highly dependent on getting kills, because if you don't, you can't even fall back to jungle with your first skill. If you get a somewhat decent timing for your level 6, grabbing your ult should definitely help in ganks.

In either case, be prepared to fall off hard. Your teamfight contribution is more or less a 300 damage nuke with a 10 second cooldown, and whatever support items you've managed to scrounge together. Probably one or none. However, you can still splitpush even as a support, because most of your pushing powers comes from having level 4 in Spawn Spiderlings, not any specific item. In general, the items are only there to make you tougher to repel (well, necro and cuirass help your push too). In this scenario it's best to build a blink because it makes you almost impossible to catch. You can either set yourself up in the enemy jungle and safe lane, or their secret shop area which grants you access to both mid and hard lane. The latter option allows you to be more annoying, and may get you a courier snipe or two. However, the safe lane has more farming opportunities. Then again, if the enemies actually want to farm their own jungle you might not want to be there, unless your team has good gankers in which case you can easily spot opportunities.

Because I love playing Brood so much, I'll try both approaches.


So I did try this both ways. It turns out you can jungle by starting off with Spawn Spiderlings, but you need to be extremely careful. Killing of your spiderling can be a disaster. I did start with the easy camp, which some pulling supports might object to, and then proceeded to the medium camp. As long as you have one control group for spiderlings and another for spiderites, it's fairly easy to keep the neutrals hitting your spiderites, allowing your spiderlings to spawn more babies. If all else fails, tank with your hero - that's how you do the first camps. After getting level 3 and soul ring it gets easier. The problem is that jungle Brood is very greedy, as you can't do much until you can lay down some webs that actually reach lanes. She's no Enigma, but should farm at a decent speed.

I also tried the laning approach in a bot match but... the bots are idiots. They instantly forget you were ever there as soon as you turn invisible, which makes them very easy to kill with just an orb of venom. Once I got to level 2 the bots were just dying constantly. While dominating helpless AI with one of the silliest support concepts ever was fun for a while, the experiment didn't give a very good idea of how it would play out against someone with an actual brain. You're definitely not gonna kill anyone... I think. I did play one game of All Random with Broodmother the other day. I was mid 1v1 vs Shadow Fiend, and decided to try out a 0-2-2 build. The SF player was probably either very new or very drunk, because he constantly came too far into the river, allowing me to kill him repeatedly with nothing but right clicks and the slow from Incapacitating Bite. So, I guess it can work...

Because you *can* get something out of the jungle and apparently can kill people with right clicks and Incapacitating Bite, I'm going to give support Brood a rating of "poor" instead of my initial "trash".

Dota 2 Next Level Meta: Support Queen of Pain

Screw the alphabetical order! The format of this post will also be slightly different because I have already been playing support Queen of Pain in several of my solo queue games recently. Basically it's less hypothetic and more based on personal experience. My initial estimate for QoP: completely legit as a support and I do think this is the case. In fact, I think there was a team in one of the TI5 qualifiers that ran support QoP with a 1-1 record (Chinese maybe?). She's a very popular hero at the moment, and I'm expecting to see nerfs in her future... but at least until then, this post should be valid.

Hero analysis

Much like one of the absolutely worst heroes to play as a support, QoP also comes with a blink ability. However, unlike Anti-Mage, she actually has other abilities that are useful without items. So, she can actually make use of her blink. Her first skill, Shadow Strike, has been buffed and buffed over multiple patches, up to the point that it's now a really strong level 1 spell with 200 magic damage and a 20% slow. Admittedly, it takes 15 seconds to get all of the damage, and the slow is deceasing (meaning it's reduced over time). The spell also scales in almost every aspect: damage goes up to 425, slow to 50%, cast range increases and the cooldown decreases to 4 seconds. It's an incredibly strong spell for harassing in the lane. While it's definitely no Venomous Gale when it comes to killing people, it does do way more damage at level 1 and has a lower cooldown and mana cost - especially in relation to QoP's high mana pool.

QoP's blink has also been buffed recently to 1300 range at all levels - that's almost double the prior level 1 range. Needless to say, this makes it a very potent mobility spell in all stages of the game. The cooldown is still significant at earlier levels, so blinking in should generally be avoided if you're not fully aware of the situation. Nevertheless, it does give her an insane initiation range: level 2 QoP can Shadow Strike an enemy up to 1750 units away - that's only 50 units shy of normal day vision range. A queen has no use for smokes (well, technically you do if you need to pass wards). However, if at all possible, it's better to walk in and save your blink for escaping - she's made of paper after all - or finishing a kill when the slow runs out.

Third skill, Scream of Pain, is her main damage source and farming tool when played as a core. Low cooldown AoE nuke that hits invisible and fogged units. As a core you generally  max this skill first (either by level 7 or level 8, if you want a second level of Shadow Strike for laning). The obvious downside is that the AoE is centered around her and it's not that large. Did  I mention she's made of paper? Spammable spells are also often difficult for supports to make full use of because they generally lack mana regen. As a mid you get mana from a bottle and QoP's amazing rune control, as a support there's no such luxury. Still it's not a bad spell to have, as it allows you to deal damage later in the game, and also accelerates your farm during downtimes. The fact that it can't be juked also makes it occasionally useful for finishing off a kill.

Her ultimate is as potent as ever when playing as support. Because she also has a built-in blink, getting into position for a good ult is not particularly dependent on your items. The latest buff made her ultimate do pure damage, which makes it much more useful in the late game, and also has certain situational advantages like blowing up Huskar effortlessly. The only way in which Sonic Wave is somewhat worse on a support is the fact you're leveling up slower, and you won't have access to an early Aghanim's upgrade (unless you're snowballing really hard). In summary, her skillset is extremely good for dealing magic and pure damage, and gives her possibly the best mobility in the game. The biggest downside is her lack of any sort of hard disable or silence. As an intelligence with good growth, she has little to no mana pool problems. Finally, her right click also has some bite with a slightly better than average base attack time, good range and decent damage. That, and her blink generally allows her to get off a couple of extra right clicks in a chase.


Building aghs on QoP as the first major item has gone up significantly since they made Sonic Wave's damage type not just pure, but also spell immunity piercing. The scepter is gotten because of the immense cooldown reduction it provides - with 40 seconds, the spell is almost guaranteed to be up in every fight. Considering that even level 1 Sonic Wave hits for 290 pure in a huge AoE, that's simply too good to pass. Even as a support, the default item build should include Aghanim's Scepter as the first major items. However, I usually don't straight out rush it. My most usual build when playing QoP as a roaming support is to start out with as many Mantles of Intelligence I can get away with - after getting a pack of Tangos, two Clarities and whichever support items I need to buy. So, ideally, two. Inspired by the short time trend of the Blade Mail Prophet (NP build with three or four Null Talismans and a Blade Mail), I build two or three nulls in the early game.

The point of all these nulls is to make your right click even more potent, and also to boost your early game mana pool. The rationale is much the same as it was with the Prophet build: you have a lot of mobility thanks to your skills, so there is a lot of benefit in bringing as hard a right click as possible with you. That, and aghs buildup is shit for damage. Only one component gives you 10 damage, and it costs almost the same as two nulls. Two nulls would give you 174 HP, 18 damage, 168 mana and almost 1 point of armor (and 6 attack speed, yay). Sure, eventually some of the nulls will be sold, but a couple of hundred extra gold for a huge early game boost is definitely worth it. However, if you are stuck sitting in a lane for some god-awful reason, then the nulls don't really do much. I usually try to go for three nulls; however if I need to spare a slot for wards, then two is usually better because the third one would need to be sold so early (to make space for your Point Booster).

I usually upgrade my boots after I have all my nulls, and the upgrade of choice is always Power Treads because it gives you the most right click DPS. The extra speed from both Tranquil Boots and Phase Boots is kinda wasted on QoP because of the blink. After treads it's usually aghs time. Occasionally something else might be needed though, and in my experience it's usually Eul's. It's not ideal, but if your team severely lacks anything to cancel channeling with, you are simply going to need it if you're against something like Enigma, Crystal Maiden or Witch Doctor. It's also the most inexpensive way to save yourself (or a teammate) from a Spirit Breaker charge. It can also be used to remove a pesky silence, and the mana regen makes farming with scream a lot more affordable.

Finally, I would advise against getting Orchid on support QoP. Well, I generally advice against it on pretty much all supports. Orchid is a snowball item, and cores who build it generally build it first *because* they can get it fast enough. It's a powerful item in the early mid game because of the solo killing potential it provides. However, once BKBs start to come up, it loses much of its usefulness. Because supports aren't known for being rich, you don't want your one major item to lose its potential as the game goes on. This used to be a problem with aghs on QoP too, but now that it pierces magic immunity and does pure damage, its impact is guaranteed throughout the game. So, when it comes to picking your second item, I'd personally default to a Hex, but just grab whatever feels most useful for the game at hand. There's no shame in having just aghs and a couple of nulls though, they keep you perfectly relevant throughout the game.

Game plan

As a support, Queen of Pain should be played like a support Bounty Hunter. Where BH has his invisibility to make his ganks potent despite his lack of early damage and disables, QoP uses her blink for the same purpose. Generally speaking, QoP brings more to a gank than BH does, especially on very early levels. Shadow Strike is a decent slow and does a good amount of damage whereas level 1 Bounty only has a small damage bonus from Shadow Walk and the slow from an Orb of Venom. The fact that QoP is ranged also allows her to generally get more right clicks in a gank. However, on level 1 QoP has to walk in as you really can't take blink first. On level 2 they both have a slow, but QoP should be able to deal more damage. QoP needs to be more mindful of observers, while Bounty is more concerned about sentries.

If you are familiar with roaming as Bounty or Riki, playing a roaming support QoP should be quite familiar. You simply rotate around to lanes that are gankable, and you kill stuff. Rinse and repeat. Most of the time this will be either mid or your safe lane. However if the enemy has a greedy afk jungler, you can also try to pick them off solo. To do this, just blink into their jungle from a spot where you're least likely to be seen by a ward, then try and use high ground vision advantage to find them. If they're low - which is fairly common for greedy afk junglers - should be an easy kill. Whatever you do, don't lane. Laning support QoP quite simply sucks. Yes, you have decent harass but if they actually go on your carry, you can't do anything.

Because of this, it's generally best to play support QoP in lineups that actually have gankable lanes - or against a greedy jungler. If your mid and safe lanes have no killing potential, your chances of getting anything out of roaming are very slim. A greedy jungler changes the game a bit, because you can kill them, and you can also turn your offlane into a 2 vs 2 by showing up, and with a decent offlane hero you can actually kill either their support or carry - or at least shut them down for a bit. If your team has no killing potential, it might be better to dual offlane than go safe trilane, just so you can at least harass their carry.

By level 5 the build should be 3-2-0, i.e. you should aim to max Shadow Strike first before taking any levels of Scream of Pain. It simply does more for you when playing as a support, and it's much easier to sustain with your mana pool. When you hit level 6 your ganking potential shoots through the roof, especially if you manage to get there in decent time through some early kills. For a while you should have a guaranteed kill every time your ult is up. After maxing Shadow Strike at 7, there are two options: max blink, or start getting levels of scream. Maxing blink gives you insane mobility, but your damage starts to taper off. If you have the mana pool, it's probably smarter to go for the scream. By level 11 you would then have 4-2-3-2.

Another possible route is to leave Shadow Strike at 3, which would give you 3-2-4-2 by level 11. As usual, level 11 skill builds are extremely important because the experience requirement to get to 12 jumps significantly. The 4 second cooldown is really ludicrous though. You should judge all this based on how you expect the midgame to go. If there are no big teamfight heroes, getting high mobility and a spammable slow is stronger for small skirmishes and pick-offs, whereas if teamfights are to be expected, higher AoE damage is required of you. Even if it's just one scream after your ult - the odds of you dying after getting into scream range are quite high *unless* you can walk in and blink away. However, it's more likely that you need to blink to get a good ult off.


Since I have been playing this in my solo queue games a few times, the data is a bit better than usual. Generally speaking, what I said about playing like roaming support Bounty (or support Riki) pretty much holds true. All of the various roaming heroes have their advantages, but their base gameplay is the same. If you succeed in ganking lanes, your team is set up to snowball, and you will get your necessary items with decent timing despite having very few last hits. If you fail, there isn't much you can do afterwards. Basically you're reduced to a blink-in suicide bomb, getting off your ultimate and one scream in a fight. After this you either don't get a whole lot done, or just die. In my games I generally didn't die all that much and always did decent damage, regardless of whether we won or lost.

The damage from your ult doesn't easily become negligible - the game needs to go very late for that to happen. There's also always high utility in having a blink that cannot be disabled. However, you're only providing damage, and later on disables become increasingly more important. So as a hero, QoP doesn't really offer any way to counterplay the opponent unless you get items - which is why orchid and hex are such popular items on her. The odds of you getting a hex after getting an aghs are very small, assuming there are at least three heroes other heroes in your team who want farm. In some lineups it's better to give more farming priority to an initially support QoP if at least one core is more dependent on levels than items. Blink into hex initiation is simply much stronger than what many heroes can offer - even just blink into Eul's can be very potent.

All that said, I don't think there's any particular faults with QoP that wouldn't be present in some more common support heroes. I initially estimated her to be fully legit as a support, and indeed she is - especially in this patch. She might not be the ideal support for all games, but in most games she can do a lot just by roaming. That, and I've always had a lot of fun when I've tried support QoP. Being extremely active throughout the game is often rewarding, and I think it also teaches you a lot about the game.

Friday, June 5, 2015

Persona 4 Golden - part 2

While part one of this two post series was all about Persona 4 Golden as a story, this part covers it as a game. Ultimately its success doesn't come down to anything drastically surprising. It simply has damn good implementations of important game design principles.

1. Day by day

Here we ask a simple question: what enables Persona 4 to be a game of meaningful choices? Answer: time, or more specifically, time as a limited resource. Most things in the game require the player to spend their precious in-game time. These choices are made on a day-by-day, basis, with most days taking just a couple of minutes. Occasionally the game hijacks some days away from you due to scripted events and they usually take more time but that's fine too - these events move the plot forward after all (or in some cases they are just hilarious character buildings events). What's important is that no matter what you do, your days with the game are numbered. If there was a way to somehow get more days, the day-to-day routine would lose its meaning. At the same time, time is the most elementary resource in the game: it can be converted into anything else, but nothing can be converted into time.

To put all that in another way, it's not possible to grind to undo choices. In most RPGs if you misspend your resources, you can correct the mistake by farming more. In P4 you can do this for money and experience, but every time you make a choice on how to spend your time, there is no taking back. There are ways to save time of course - most of these available to players who have high understanding of the game and a strong sense of overall strategy. For instance, being able to complete dungeons in one day each requires some planning and preparation, but pays off by freeing days for other uses. Overall, the entire game is one huge optimization problem, where individual parts are small but interconnected. Fortunately you don't need to solve the entire problem before launching the game. Figuring out your schedule can be done as you go which is what ultimately makes the game interesting.

Of course, there are other ways to play the game. In truth, the first playthrough is more likely spent exploring what the game has to offer and getting familiar with stuff. However, given that this was the fourth playthrough for me, solving the optimization problem through day-to-day decisions was the biggest appeal. Whether it is for exploration or optimization, it remains a fact that the game has an extremely high frequency for meaningful choices. Considering that each day takes only minutes and most of them contain two time slots you can fill with whatever you want, you are very highly involved almost constantly. Ironically enough, I actually consider the player being more involved outside dungeons, even though they are touching the controls way less. Although you seemingly do a lot in dungeons, the stakes of actions taken are much lower. This doesn't mean the dungeons are boring. It just means your strategic, long-term decision-making is put on the backseat while in them - or you could consider them tests of how good your strategy is.

I have another example where limited resources lead to more meaningful choices when overall strategy is concerned. There is one major difference between Fire Emblem 7 and 8: the latter has random battles spawn on the world map, and also a challenge tower that can be attempted as many times as the player likes. Meanwhile, 7 is just a series of battles, each with a set amount of enemies. This means that experience is a limited resource in 7, and infinitely available in 8. Distributing experience to characters in FE7 is extremely important whereas in FE8 you can always put in more hours to correct your mistakes. The temptation to use your highest leveled characters to kill tough enemies is much higher if you can make up for the lost experience later whereas in FE7 you need to plan more carefully so that your weaker characters who need the most experience get the killing blow. Well, I did that in FE8 anyway because it saves time. Still, moral of the story: if at least one central resource is limited, choices regarding that resource are more meaningful.

This discourse gets quite close to one popular argument: are respecs good or bad (Diablo 2 is a central game in this argument). However, I don't want to go there as it is quite far removed from the topic. Just to wrap up, it's not just the limited nature of time as a resource in P4 that makes it so amazing; it's also due to the fact that choice frequency is staggeringly high in the game. One factor that also contributes that I haven't mentioned yet is that all days are not created equal so you don't have the exact same choices available each day. However there aren't too many unique days either, so for almost everything you can do, there will be multiple opportunities. So overall, the day-to-day structure gives the game a rather unique appeal and is a huge contributor to its charm.

2. Musings about chance

Time to go into broken record mode. This may, therefore, feel immensely familiar: enemies should be dangerous, while also quick to defeat. Especially in games where combat is frequent, nothing quite destroys a game like battles that drag. I want to talk about this again because Persona 4 is a good example of how things should be. Much like many other Shin Megami Tensei titles, enemies in P4 are really, really dangerous (at least on hardest difficulty). If you're unlucky or poorly prepared, they can wipe out the entire party. So, unless you want to chance it, it's best to give enemies as few turns as possible - ideally, zero. You don't exactly get to save all that often either, so getting wiped can set you back for like an hour. Saving between dungeon floors does become easier if you get a skill that allows you to go back, but if you go back during a floor, it always resets.

Typically, the entire process is fast. You either kill or disable most enemies in the first round, and usually the entire fight ends during the second round at latest. It may sound like a cakewalk, but it really isn't - it's just that battles are decided and concluded quickly. When there's hundreds of battles in the game, this is more than welcome. It's not as much about the challenge of a single battle, as it is about managing an entire floor's worth of battles (or more) because abilities that allow you to perform those lightning fast takedowns have costs, and those costs do pile up. Let's talk about two costs. The first cost is the cost of learning. In order to know your enemies' weaknesses, you have to try stuff on them either blindly or by making educated guesses. This means that encountering a new powerful foe is especially dangerous because you possess no certain knowledge of how to disable it quickly.

There's a sidetrack here. While I generally like that allies have been given more skill options through social links, some of the things Rise learns when you level up her S.Link are a bit too strong. At some point she can show you every detail about enemies for no cost, which takes away the cost of learning. Of course at that stage of the game you are probably kind of snowballing out of control anyway (see the next section). The second cost is the cost of certainty. As stated, it's not that enemies are guaranteed to kill you, they just have a chance of doing so. Which means you can take things slow and be more conservative with SP, but that means rolling the dice more. Using abilities generally means you don't have to take the chance - so, basically, you pay for the certainty of success. As a bonus you also end battles faster. This is a common design pattern in RPGs.

As such, the cost of certainty comes down to the essentials of game design: meaningful choices. When faced with a combat situation, you have to assess the stakes (e.g. how long has it been since you saved), the risk (how likely it is to actually lose) and of course the cost itself. In some cases there are more than two options, with varying costs and risks. Chance is an important part of this equation - the dynamic becomes different in a fully deterministic game - you would maybe choose between different types of costs, or instead of chance, you are betting against uncertain future (e.g. "do I need this resource more in the future"). In skill-based games, chance can be substituted with player ability (e.g. "can I pull this off"). However, in turn-based RPGs, most of the time such choices are made against the RNG. For an extreme example of how important chance management is, check my boss analysis of Digital Devil Saga's Demi-Fiend

Persona 4 also uses chance to give the player freebies. Most of the time these freebies allow you to use less abilities. The most common ones in P4 are follow-up attacks, which may occur when an enemy is knocked down. These generally knock more enemies down, except Chie's which instantly kills another enemy. P4G adds a few other similar freebies, all of which are welcome. Of course, critical hits are perhaps the most wide-spread general freebie, and in many games you can also manage your critical hit rate in some way (P4 - not an exception). While this may sound a bit like "random shit happens - I have no idea why", let me assure you that it's very welcome. It really helps in keeping battles interesting, and it does feel good to get a freebie - especially if you really needed it.

3. Option expansion

I have touched this topic several times in the past. It's a pet theory of mine that I've never quite put onto paper in full. I use it to explain why I think a lot of games - RPGs in particular - generally get easier and easier the further you are into them. At least if you have like half a clue about what you're doing. As the name suggests, the theory applies to games where the player is given more and more options as the game progresses - which, coincidentally, covers most RPGs as it's kind of a consequence of having all those character development aspects. So, in short, the available option space expands. This also means that the further we are into the game, the higher the discrepancy between choices made by different players will be. So, from the designer's perspective, the later the game goes, the harder it is to predict.

This puts the designer in a somewhat tricky situation. Not everyone utilizes the expanded option space as efficiently, and a lot of people play RPGs for their content. Therefore it's important that anyone can actually finish the game but at the same time it also very likely leads into a situation where highly optimized builds bulldoze through everything. Not to mention that - unless the system is quite simple - just predicting all the possible interactions between options is very unlikely to succeed. So, the theory states that it's near impossible to create challenges while at the same time accommodating multiple strategies and skill level. Games generally have difficulty levels to deal with these problems, but often they use some kind of numeric scaling - and as such, numbers don't actually matter that much if the player can completely ignore some of them.

You see, often it's not just that the number of options increases over time - usually the later options are also stronger. This further complicates things, especially in systems where the player has more control about the order in which they acquire their options. The Shin Megami Tensei franchise has one signature breaking point in skill development: Mediarahan - a healing spell that fully heals all allies. Before that point it may take multiple actions to heal your party to full HP, but after that point, one spell is 100% guaranteed to do so (unless someone died). Ultimately you also get a spell that heals all status effects as well, but as a leap it's way less significant than Mediarahan. This one spell completely shuts down any challenges that are designed to whittle down the party's HP faster than they can heal it.

There isn't anything particularly overpowered about Mediarahan, mind you. Many RPGs have some absolutely bonkers, broken shit that, once attained, becomes such a dominant strategy that the game might as well play itself. However, a dominant strategy is just an extreme instance of this phenomenon - it's a strategy that shuts down everything in the game. Usually, slightly too useful skills just shut down certain dynamics from the game. Generally, any options that completely negate something with 100% success rate inherently reduce the impact of certain types of dynamics to zero. It doesn't matter that enemies hit for, say, quadruple damage if all they get is the quadruple of zero due to players entirely avoiding the effects of the attack. There's a cornucopia of examples available (from various games) when it comes to skills that make actual numbers meaningless.

If you look at the toughest boss in Persona 4, Margaret, the first thing she does is to negate any immunities characters may have. This is a recurring theme with optional super-bosses. They simply have to negate some of the player's options entirely, because those options would make it too difficult to create an interesting challenge. Super-bosses are also usually designed for characters close to the maximum level, and access to all options in the game.They are often rather elaborate designs that tend to require very specific things from the player's strategy. In other words, in order to truly create a challenging boss, the designer needs to artificially cut the option space back to a manageable size. While this tends to create rather puzzle-like encounters, it's still welcome as opposed to players being able to absolutely destroy everything in the game with no resistance whatsoever. Of course such encounters should be optional content.


This post has been open so long in my editor that it's time to just release it. Admittedly I'm getting lazier with this blog, and I should probably change the way I treat games. Probably there were more things I intended to mention in this post, but as it stands it sums up the main points I had to say about the game pretty well. The thing about Persona 4, is that it's just incredibly enjoyable to play at every turn. A few sidequests aside, pretty much everything, every single moment, in the game feels meaningful. Although this post explored some dimensions, there's definitely more. There's just this sense of overarching quality that has succeeded in capturing my full attention ever since I started the game for the first time on my PS2, all the way to when I finished my fourth playthrough on Vita.

The re-release has its flaws. Rise in particular gains way too strong abilities if you level up her social link. Likewise the unique, ultimate powers of Chie and Naoto are clearly out of whack. The new very hard difficulty is a bit lazy (it's the same as hard, but you get less experience and money). The new optional dungeon is also quite stupid, and a bit lazy. The original game also had some flaws, like the fact you absolutely had to be on a second playthrough in order to access everything in the game. This is especially annoying for me because they retained this requirement in Golden, meaning I would have to play it once again... but then again, I have already beaten Margaret in the original game in it wasn't really the most challenging secret boss so not a whole lot is lost.

I hope this two part series was useful for understanding why Persona 4 is the perfect fit for me. That way it should also shed some important light on my other posts about other games. I haven't actually played that many games recently, so I have no clear idea what I'm going to write about next.

Thursday, February 5, 2015

Persona 4 Golden - part 1

I've been itching to write about this game ever since I started playing it. Let's start with a bit of background: the original Persona 4 has been my absolute favorite game since it came out and I've beaten it thrice already. There was no way I could keep my hands off from an improved version - even though it was released on a console I had very little desire to own. That did delay me a bit, but eventually I had to buy a PS Vita. I mean, I had already bought the game from a PSN sale anyway. I also delayed a bit because I knew from reviews that P4G would not have nearly as much new content and other improvements as Persona 3 Portable had. Anyway, It was nice to find out that the my most beloved game ever was just as lovable as it was over 5 years ago. I guess we are here to find out why.

Or we would have been, but the post started to get really long. Instead, this is part 1, and it discusses the game from a story/characters perspective, and the roleplaying aspect that I discussed some time ago when going through Dragon Age: Origins.

1. RPG?

This is a good continuation to my previous Dragon Age post. I have probably mentioned this in passing before: I find it peculiar how the term roleplaying game is used. As it stands, most titles in the genre have next to nothing to do with roleplaying at all. DA was one of the few exceptions, and in general it's more common to have some roleplaying in Western RPGs. However even here we have a legion of so-called action RPGs that have absolutely nothing to do with playing a role. What is generally called an RPG actually represents some other genre but has character development as a central aspect (a bunch of action-adventure games already have simple character development, but usually it's just glued on top). Japanese RPGs in particular have very little to do with roleplaying; they are linear like a railroad track, and their game mechanics are in many cases very abstract.

So although they supposedly fall under the same category, and even sound similar in many ways, there is very little common ground between Persona 4 and Dragon Age. You might think that all those Social Links in Persona 4 would be opportunities for defining the main character's personality like I wrote in the previous post. Alas, they are not. Although you are offered some choices during S.Link events, these are not really roleplaying choices. They are quiz choices that test how well you understand the other character's personality, and you are rewarded for making the right choices as your relationship grows faster that way. So the game does not only make value judgments about choices - they are also tied to explicit rewards. The relationships also grow linearly - only the speed of growth is accelerated. There is one roleplaying choice involved in some S.Links: with the girls, you can choose whether you want to be friends or lovers.

Likewise, on the first playthrough you might think there are choices involved in who you want to spend time with. Again, this is an illusion, and a result of poor scheduling on your part. In fact, it is possible to complete every S.Link within one playthrough. Therefore there exists at least one explicitly optimal path through the game where you don't need to make roleplaying choices because you basically get everything. At this point it's important to emphasize that this is not in itself a bad thing. It simply means you should make your choices from a strategical basis: you need to be able to understand the game's schedule in order to create your strategy (of course you can kill the fun by looking up a day-by-day guide). You also make strategic decisions about the order in which you raise S.Links. Beyond that there are yet other choices to make - indeed, whether you actually want to max all links is also a choice, because there are benefits in spending some extra time on other things.

The point is that, due to the nature of choices, P4 is not a roleplaying game by design - not in the sense of the word's meaning in tabletop circles - whereas Dragon Age is (at least to a much higher degree). It is however a game with meaningful choices. Indeed, the choices you make concerning how to spend your time have a huge impact when it comes to exploration and battles. All this means is that the game caters to different preferences - and these just happen to be pretty exactly my personal preferences. Almost every single thing in the game contributes to your overall strategy. This makes each individual decision interesting, as there are always strategical considerations involved. As much as I praise the game, one change would have made it better: it should have been impossible to max out all Social Links in one playthrough. Or, at least, there should be no special reward for maxing all of them.

The reason I feel this way about being able to max all S.Links is that it becomes a goal that is very irresistible. Although there are multiple ways to reach this goal, it still limits what you can or cannot do during a playthrough quite tremendously because most of your time will be contributed towards this huge effort. You can of course choose not to pursue this goal, but the game does portray maxing all S.Links as kind of an ideal result. The same problem exists, although more severely, in Valkyria Chronicles. There the way missions are evaluated strongly prefers one particular type of strategy that is very limiting and even boring to play - so the game encourages you to not have fun basically. It's not as bad for P4, because it is still a lot of fun to figure out how to accomplish that lofty goal. The game is lax enough that you don't need to follow a step-by-step guide to do so. Following a guide would totally kill it for me. For the record Persona 3 is much stricter in this respect.

Just to clarify, you definitely *can* play Persona 4 like a roleplaying game. I have some vague recollection that this could have been my approach on my first playthrough - if not in P4, then at least in P3. Roleplaying is simply not really reflected in the game's design. The main character does not really grow nor are you really presented with any opportunities to reflect upon his personality. Likewise, since choices aren't created equal, they are not exactly choices in the roleplaying sense.

Since it's not a roleplaying game in any traditional sense, using the abbreviation RPG sounds silly. I make a point of always distinguishing Japanese RPGs with the J for exactly this reason. It doesn't make it less silly that the letters R, P and G are still there, but at least it signifies the fact that there is actually nothing related to roleplaying in this game. To me at least the J in JRPG is a clear genre indicator, instead of place of origin.  It quite clearly states that this is not a roleplaying game. It's more like a mesh of tactical battle game + anime series + RPG style character development. The term RPG is there because early on these games shared a lot with traditional fantasy roleplaying games. Since then, tabletop RPGs have evolved way more towards roleplaying, while JRPGs have evolved into some other direction entirely. That's my best guess anyway why the letter combination is still used.

2. Real people

Another topic in my last post was about relatedness of virtual characters. Persona 4 is an interesting game in this regard as there is a sharp contrast between the protagonist and all his friends. It's not much of a surprise that characters in the game are generally super-relatable. The combined length of all cut-scenes in the game is way beyond normal - almost on an entirely different scale really. For comparison, a normal 26 six episode anime series runs for about ten hours. I didn't count, but I'd wager this would be at most half of all the cut-scenes in P4G, so it's at least in the "one season of Buffy" territory - at least; some sequences in the game are as long as movies. If you want to write an entire, huge anime series into your game, it needs to be good on TV standards or the entire thing becomes a horrible yawnfest. So falling back on game standard for your characters and their representation is definitely not going to cut it. 

Of course it also goes the other way around. The more screen time characters get, the more relatable they can appear. However, this is definitely not a 1-to-1 relation - simply giving characters more time to talk doesn't necessarily make them any more relatable. In the last post I used Mass Effect's characters as examples of poor presentation - despite given a whole chunk of time to talk, they never seem to become much more than audio book autobiographies. Another trope in video games are characters that are walking encyclopedias or plot explainers. They never really react to anything on a personal level. All in all, they appear rather emotionless, and consequently scenes where they do pour out their feelings come out as paper thin. Dialogue can be entertaining even in these scenarios (e.g. the plot can be interesting enough to warrant a few hours of explanation) so it doesn't mean these games are automatically bad - but you never really feel the characters.

Characters in Persona 4 dick around. A lot - so much you'd almost think they are just high schoolers playing detective. Oh wait... that's exactly what they are. They talk to each other constantly - often just to shoot some random crap - and they react to things that happen around them. Even serious discussions about the case or studying turn into food discussions or whatnot eventually. They make fun of each other in a playful fashion, they support each other and really just bounce around from topic to topic rather randomly. Most of the game's events aren't honestly about the plot at all, they're just all kinds of school events that usually end up weirdly for our heroes. Sure, they abide by anime cliches a lot, but nevertheless, beneath every set of bytes there is a person. Not only has a lot of effort been put to writing and designing each character, just as much effort has been put into showing off. Indeed, there's a ton of events in the game, from small dialogue scenes in the morning to the massive movie length plot ones. 

Overall, dialogue writing is just incredibly solid. Sure, maybe it's not on the godlike level of Gilmore Girls and not quite up there with Buffy either, but there's a coherent personality behind every line. I had all sorts of warm, fuzzy feelings as I followed the characters grow together as a group, and the game's humor worked for me. Of course it's a bit of tropefest and if that bothers you, then enjoying the game will probably be more difficult. However, if we disregard the content itself and just focus on narrative methods for a while, there clearly are a lot of things done right for a game with a linear narrative. There's this massive argument whether games should sway away from cinematic narration or not. I personally don't mind the lack of interaction with stories in my games. However, if you want to go that route, you should follow the writing standards of your source medium - and for your typical RPG, the equivalent video narrative is often a TV series. 

3. Unreal people

Spoilers. Minor, but spoilers nonetheless.

In sharp contrast to all the other cast member, the protagonist is enigmatic. Like most games in the Shin Megami Tensei franchise, Persona 4 employs the silent protagonist paradigm. Although he has a voice, the main character has no voiced lines - indeed, anything he ever says is selected by the player from a handful of options. Yet because of the game's highly linear nature, the player can only affect nuances in his character. No matter how badly you do in your exams or how few Social Links you attend to, in the end all your investigation team friends will still revere the main character. So although you can choose to fuck up as much as you like, in the end the protagonist is still written to fulfill the role of mr. perfect. Indeed, if you play well by any of the game's standards, the main character will come out as a superior being. As being mr. perfect is coded into the game's reward structure, I'd consider the main character pretty much written in stone.

I don't even think of the MC as a real person per se. If you look at how other characters in the game grow - especially in Social Links - they credit most of their change to the MC. However, he doesn't really do all that much. To me it makes more sense to think of him as some kind of a mythical paragon or a mirror of the soul. It's kind of consistent with his special power: he can literally assume multiple personas to fit his needs - or, more likely, the needs of his friends. Not only does this allow him to defeat anything thrown at them - it also allows him to be exactly what his friends need to overcome themselves. The game even hints at this: if you are carrying a persona of the friend's arcana, you get more relationship points when spending time with them. As the MC's understanding of a person grows, so does his power in relation to that arcana - and he can act as a better mirror, allowing people to face themselves more and more.

This is kind of a philosophical meta level to view the MC. I do think everyone - including the MC himself, at least consciously - believes that the MC is just a normal human person. This is why the only thing that really seems a bit off is his ability to be awesome at pretty much everything. Whether there ever was a true person inside or not is debatable; however, I do think that what everyone sees is just multiple reflections of people's hearts - and in a way, of the player's wishes. This is probably even more pronounced in Persona 3, at least to a Western audience. The MC of P3 is portrayed quite clearly as a sort of messianic vessel. It's not about just saving the world in either game. After all, the catastrophe in both games is ever-looming, secretly desired by mankind. This is why it's not enough for the MC to be a hero - he needs to be a paragon that truly changes people. On a higher level, P4 is about making mankind desire the truth.

So. How is this relevant? It's a viewpoint that makes all the pieces about the game fit. If you think about the MC as a roleplayable character, you are bound to be disappointed by the lack of possibilities to truly define him. Likewise, he is not a scripted character either - he barely has any lines. However, if you accept my rambling about how he is a presence more than he is a person, the contrast between him and the rest of the cast kind of disappears. In a way it could be said that he is a direct extension of "you" - the player - in that he is not truly present in the world as a character, yet he has more influence over the world than anyone else. The same thematic repeats with the main character in most Shin Megami Tensei games where you get to name the protagonist - with the important exception of Raidou Kuzunoha who is a person with a background.

It's worth remembering that as a genre, JRPGs generally position the player entirely outside of the game's events. The game may have a protagonist on paper, but ultimately it's always about the entire group, and the player can only follow from the sidelines. The difference between P4 (and others) and most of the genre is the fact that the player is observing from inside a character - but the dynamic does not change: even if the player is positioned inside a central character, their only purpose regarding the story is to follow and observe. That is all.


Since this threatened to be become yet another massive post, it's better to stop here for now. To wrap up: I absolutely love the characters in Persona 4, and they are one of the key reasons why I like the game so damn much. The story in itself is not massively impressive but it does create a better environment for the characters to goof around than the one in Persona 3 (I think the plot itself was better in P3 actually, but it's not as enthralling to follow). Sure it's a long game (actually it's effing huge) but there is something happening almost constantly. On top of that, there is the ever-present philosophical meta level to think about. This is a common trait in Shin Megami Tensei games, and I would not be surprised to find it among the reasons why people praise them. Still, what really makes P4 possibly the best game ever made, is first and foremost a matter of actually playing the game - and we'll go there next time. Until then...

Thursday, January 29, 2015

Global Game Jam 2015 - Mahou Shoujo Idol Strike!

It's that time of the year again! This was my fifth GGJ in as many years. Although I have gotten lazier when it comes to other jams organized by our game dev club, I have at least consistently participated in the biggest and baddest. Results may have varied... Anyway. Let's get into this. This is divided into a few sections - the first one is a bit more self-reflective, and it's about game ideas in a jam environment. The rest are about this year's project - which I'm sure you are dying to read about after reading that title.

1. About ideas

Let me just start by reminding that this section is kinda self-reflective and massively opinionated. So keep in mind that the most important thing when it comes to game ideas at jams is to do whatever the fuck you want - you are there to have fun after all. However, I'll be a bit more philosophical. Game jams are learning experiences and to that end, you may have some kind of a learning goal. Maybe it's to become better at your chosen profession, or maybe it's to try something new. I'm a decent programmer, but when I go to jams, I go there to work on something that has a fresh design. Therefore it is incredibly difficult for me to understand why anyone would want to spend their weekend making an underdeveloped version of a game mechanic that has already been developed by actual studios or indie developers. I do realize of course that it may be an interesting technical challenge just to see if you can get it done. Likewise if it's your first jam, simply getting any game done at all can be exhilarating.

It may be a pessimistic to way to look at things, but really, after too many jams, I seriously don't expect a jam game to be good. Like, at all. If your goal is to make a great game in 30 hours, prepare for a lot of failed jams. For me, personally, the measure of success is whether I got to try out something different or not. Probably this is why I have worked on my own idea in every jam so far - it's usually some mechanic that simply intrigues me too much. That, and a lot of the ideas of others have generally failed to inspire me. Again, this doesn't mean they are bad (indeed, the outcomes have often been more fun to play than mine) - I just feel like their owners have different goals for the project than I have. That, or I don't really see the game from the pitch. I think I actually found a way to work on the latter this year. Normally I would just think about some ideas by myself, or even leave the entire site to get some food or whatever.

This year, I was originally supposed to lead an "idea hotspot" - kind of a brainstorming table. We kinda failed at organization tho so we ended up not having them. Instead I just rotated around, asking people to throw their ideas at me and gave them some additional questions to chew on. I found this a nice approach actually. It gave me more insight into what others were thinking, and also may have helped them get a bit better grasp of their concepts. So this year there actually were a few other projects I might have joined. I'll have to admit, I kinda cheated this year. Literally three hours before the jam I was struck by an obnoxiously sticky game idea and I was unable to get rid of it. It prevented me from coming up with any new ideas of my own during the brainstorming time. I guess talking constantly about other people's ideas also kind of hindered my own thought process. So, anyway, after a lot of consideration I did pitch that idea... I felt a little guilty about it, but at least it was an idea I got on the first jamming day. Even if it was before the event started.

There was another reason too. Let me sidetrack a bit to say that the theme this year was absolutely fucking horrible. "What do we do now?" is a non-theme if I ever saw one. Either you come up with some incredibly bonkers artsy crap (not that I mind, I like those), or, well, you can pretty much encompass every single game ever made within this theme with very little mental gymnastics. The games made at our site this year didn't really connect in any way. It was just literally a bunch of random games - not take anything away of course, some of them very really good. Anyway, I wanted to bring this up because - like every single idea in the universe - my sticky idea from a few hours ago actually fit this non-theme. It even included the plural "we" in form of four characters. It still felt a bit stupid to fit an idea into a theme, not the other way around. So yeah, I decided to just throw it out there. And for the record my pitch was effing horrible.

I generally fail to attract workers to my cause because my games are described through their mechanics. I guess that's a bit unattractive, and they don't sound all that wild on paper. It's actually quite hard to see an abstract mechanic if it's being described to you - you can try this by reading a board game rulebook without access to any game components. This time my idea was inspired by two sources that typically get reactions one way or another: Asian pop music girl groups and the "magic girl" manga/anime trope. So even though I did a terrible job with my rambling pitch, I got the interest of one artist and one sound guy (who, curiously enough, hated j/kpop). Coincidentally that was just what I needed, so once again I ended up developing my own idea but this time with a team. So, just a casual takeaway: even though I mostly care about mechanics, I shouldn't rely so much on appealing to everyone else's inner designer. Instead, I should work more on sticking a crazy theme on top of my mechanical idea.

2. The game

I got the idea entirely from listening to kpop. More precisely, I was listening to Kara at the time and my tired mind wandered back to some recollections of their music videos. Suddenly I found myself wanting to make a JRPG battle system where all characters would move as a group, abiding to a dance choreography. In my mind it was more like a tactical RPG, even though the pop music connection might make most people think rhythm game. However, for me the core concept was careful selection of dance move patterns for the group as a whole. I also played around with ideas about having kpop-inspired roles for the characters (leader, lead singer, lead dancer etc.)  and the fact that the singer alternates frequently. Again, as my mind was already racing towards an RPG, I needed some connection between that and kpop. Probably as an indirect result of exposing my mind to too much Persona 4, it went straight towards "witches who use pop performances to fight evil" - there's probably a legion of animes about that already.

Since it was to be made into a game jam concept, some chopping down had to be involved. I kept the essential parts: there would be a group of characters and they would move in unison; each turn the player would choose from a handful of available patterns. The girls would need to move over specific tiles to activate their magical powers - this was intended to form the puzzle element of the game. Ghosts were added to the concept too. In the first iteration, the concept was like this: the girls would try to avoid ghosts while picking up power-ups; after doing so, they can cast magic to banish nearby ghosts. The interactions were quite simple, which made the basic game code very quick to write - a boon if you want to work on complicated mechanics that need a lot of rethinking. Most of the design complexity came from juggling two things: what happens on the board, and what moves the player has available at each moment. A bit of a rematch with problems I had when working on Pulselight Steampunk.

For once I even came up with a relatively straightforward way to make a tutorial: the first set of actions was a preset, and the player would see the key interactions by performing them in the given order. So at least in theory you could just tell players to press 1, 2 and 3 in succession and watch what happens. In practice the game is a bit harder to understand because the player has to follow four different characters at the same time. That in itself is fine, and gives the game a healthy dose of cognitive challenge. It became a bit problematic because for most of the development time the game felt insanely random. Only four actions were available to the player at a time, while there were like 40 actions total. The chances of having only useless (or worse) actions was staggeringly high. It didn't help that everything spawned on the board in random locations. Often, after just a couple of moves, the player was stuck with no way out.

Figuring out solutions for this problem became the essence of my game jam this year. I quickly changed some of the interactions; primarily, the girls were now able to defeat ghosts by running into them in addition to casting spells, and the ghosts could only spawn at the borders. This called for a new lose condition, and I ended up with "defend the center". So the ghosts would try to conquer the center of the board. I also fiddled around with how moves were offered to the player. I divided them into a few different categories and then made the game rotate between categories, basically guaranteeing different kinds of actions regularly. The end result feels mostly like a game where the player can actually control stuff. Still, the feeling of getting screwed over by the game is too frequent. I'll outline some next level ideas to improve the game in the next section. Before that, just a few words about its development.

Our team of three people worked quite well. Which, of course, is quite simple to do when everyone has a clearly specified set of tasks. The game's core was also really fast and easy to implement - it was pretty much done in like four hours. This is something that I highly prefer because it leaves a lot of time for fiddling around with design concepts. I also had time to code all kinds of luxury functions like support for animated visual effects. We were initially going to include a tilted perspective instead of a straight top-down approach, but improving the design took priority in the end. That, and one arcane bug that I managed to include in the code. Despite locating and fixing two different instances it, it's still present somewhere in the code. As usual I only have myself to blame for that. It's just one of those "what could possibly go wrong" things when writing game jam code.

I have found CraftyJS to be an amazing library for jam development - if you have learned it beforehand that is. There are two reasons for this: its component-entity model, and its event model. Crafty keeps track of all created game objects, and allows the developer to query them based on their components. Which means there is never a need to keep track of references, because you can always fetch them from the core object. I hate passing references so much in object-oriented programming. Passing a new reference down a few function calls already requires modifying code all over the place. Using events handlers over method calls has the same advantage. You just register a handler, and then can trigger the event from wherever you want instead of - once again - messing with annoying shit like references. Overall, not giving a damn about references speeds up development and makes it generally more relaxed.

I do think there is a downside to sticking with CraftyJS though. It only does 2D, which is a bit of a hindrance these days. Not that I'm particularly interested in making 3D games, but I should at least be able to do so just in case. Of course another problem with the library is its horrible documentation, which makes it hard to learn - definitely not ideal if you work with other programmers. So yeah, one day I should really learn how to use Unity3D.

3. Idol Strike! Next!

After the jam I thought of some ideas to make the game feel less random. Having too many random things in jam games is one of my personal faults. It's just really tempting to generate stuff randomly because actually designing levels takes a lot of time and it's not exactly as exciting as creating new game mechanics. However, I don't really see a way to make this year's game better without seriously chopping down on the randomness. Furthermore, if it were to find its way into its original role - to become a battle system for an RPG - it needs to incorporate way more customization options for the player. First of all, instead of being an endless score attack, the game needs to have clearly defined levels with specific challenges. Moreover, things like power-ups should be in fixed locations to incorporate any sort of actual planning into playing the game. Probably the same should be done to enemy spawns.

Something definitely needs to be done about randomizing available actions too. I want to stick with the idea of dance routines on this one. Instead of offering random patterns, there would be four different dance routines, each with its own fixed sequence of moves. Each turn, the player would choose one dance routine to follow, and could then choose to follow another on the next turn. The key here is that there would be a way for the player to see ahead at least a few moves or the entire routine (which could actually be just four moves actually, and then loop). To encourage further planning ahead, there could be some bonus for sticking with one dance routine for an extended period. Overall the key to making this game better is to a) provide the player more reasons to try to move to specific parts of the board; b) provide the player better means to get the girls to go where they want them to.

This way would eventually turn it into a puzzle game where the player would try to complete levels. Once the basic mechanics are neatly fixed, the concept can easily be expanded. More types of enemies are a no-brainer. Likewise, there could be more than one spell available in the game. For instance, there could be other banishment patterns than just the present "surrounding tiles" approach. After that, it's not a huge leap to make the girls individual by giving them different advantages. Thinking about extensions too much is generally detrimental though, so we might as well stop here. This section was just a short demonstration how exploring a game mechanic in a game jam game can lead to further discoveries and may, in the end, be a more valuable result than a more complete, more conventional game would have been. Contrast this to my "easiest" jam game Umbrella Dream (simple platformer) - sure it was easy to understand and kinda enjoyable to play, but I gained nothing at all from making that game.


So, there's one more game jam under my belt. Once again I worked on something that deviated from standard genre conventions and as a result had my hands full trying to juggle everything into a game. Success in making an actual game was partial. The jam itself on the other hand was fully successful. We were able to bring a new mechanic into light and see how it worked - which it did. This time the game around the mechanic just fell a bit short - but that is also a success because I was able to pinpoint how it went wrong and at the same time understand better why I've experienced similar failures in the past. That said, the chances of actually continuing development are kinda low based on prior experience. I do consider this particular project to be among the more interesting ones in that regard though. Nevertheless, just thinking about game experiments I have made in a reflective fashion at least keeps me in the right mindset.

Anyway, I think that's enough about this year's GGJ. It's time to get back to the side of games I am more comfortable with: playing them, and writing about them. Thanks once again to Stage for setting up the jam site, and of course to everyone who was there contributing to the atmosphere. This year's jam site had just enough space and people filling that space.

Friday, January 9, 2015

Dragon Age: Origins - Part 2

It's finally time to continue where we left off some months ago. After bashing Dragon Age from several angles, we were wondering why on earth I chose to plow my way through the game. This post answers that question. Once again a lot of the things I'm going to say could be said of Mass Effect. I will be talking about my character a lot. Cath was a female city elf rogue. I wanted to play a rogue as it has been a while since I have played one; I wanted to play as an elf because for me it's always either elf or human, and once again it had been a while; I chose the city background to give my character a bit of cynical and practical flavor; finally, I chose female because that's my usual preference when choosing a virtual identity.

1. The roleplaying appeal

This is pretty much the main talking point of this post. I want to use an interesting description by James Paul Gee about the relationship "real person as a virtual character", or in this case, "me as Cath", as the basis of this discussion. The key point is that this relationship has three identities, not two. The first two are relatively straightforward: obviously there is myself as the player, and there is Cath as the character - we both bring possibilities and limitations to the gameplay experience. For myself, this means my mechanical and cognitive ability to play the game; for Cath this means her stats and other things that describe her in the game world. It's the third identity - the projective identity - that makes roleplaying games interesting. This projective identity is my interpretation of Cath's personality. It is not simply me, nor it is simply Cath - it's more like me making decisions for Cath based on what I know about Cath. If that doesn't make all too much sense, worry not - just read on.

This is bread and butter for those with experience in tabletop roleplaying games. You play as a character and in most cases it's not yourself. However, neither is she or he completely fictional - the end product is a combination of both the player and the character. The same can be said of acting. What you see on the screen is not the actor nor is it the character - it's always the actor's projection of the character. A different actor would portray the character differently - within the freedom they have been given. This freedom to interpret is also highly important when we talk about digital games. In tabletop RPGs the degree of freedom with which players interpret their characters is huge and they also typically have more agency in writing their character. The game master typically sets certain limits that are necessary for his campaign to work but beyond that, characters are created by the players themselves. When the game master and all the other players are replaced by bytes of code, upholding the illusion is much more difficult.

The ability to uphold a strong sense of roleplaying is BioWare's strength. This sense is particularly strong in Dragon Age, much more so than it was in Mass Effect. What I mean by that is, the game inspires me to think about things from Cath's perspective by providing me the freedom to do so. One important part of this is the way choices are framed - or rather, the way they are not. All kinds of value judgments about the choices are left to the player. Cath's companions did often chime in with their opinion, and that is fine - more than fine actually, because they *are* people, and people have opinions. What's important is that the game itself did not. Mass Effect has this problem with its bipolar morality scale - the game is disrupting my roleplay by explicitly telling me which actions are "good" and which are "bad". Note, again, that games definitely can tell things like this, but they have to do so implicitly, by showing consequences.

When choices are framed in one way or another, the decision-making drifts away from the projective identity to the player identity. This happens because the game is telling the player something the character cannot know. The extra information can cause conflicts that break immersion, especially if there are explicit game-mechanical consequences involved. For instance, if I - based on the projective identity - make a choice that I (the projective identity) deem "good", and then the game explicitly tells me that it was actually "bad", and punishes me for it (by moving my character to the wrong direction in a morality scale), it is severely inhibiting my ability to roleplay. The conflict here is that my character thinks the choice was "good", but me as the player is given information that it was "bad" and as a result my character is also now more "bad" by a bit even though I had her projected as "good". As you can see, this is a huge problem that comes with explicit morality scales.

On the other hand if choices aren't explicitly labeled and outcomes are only revealed later through consequences, it just becomes another opportunity for reflecting upon the projective identity when things end up going south. Choices can also inhibit roleplaying if they are too limited. This is the more difficult thing to avoid in digital games, because you would need one hell of an AI to truly abide to all sorts of projective identities. The general problem is that if none of the offered choices feel satisfactory to the projective identity, the player's agency in constructing that identity is taken away and put into the hands of the game's writers. Dragon Age generally upheld the illusion that I was able to make choices that seemed sensible to Cath. At the same time, choices made during the game - especially choices concerning companions - contributed to the growth of Cath as a character. This is no small achievement for a digital game. It is, indeed, more than enough reason to enjoy the game despite its shortcomings.

It is worth noting that while being able to create Cath myself (partly anyway - her background had to be chosen from a limited number of options), it is not necessary to have free character creation to provide a strong roleplaying experience. Even if the protagonist is pre-written, a game that supports roleplaying still easily allows the player to develop a strong projective identity. Case in point: The Witcher - especially the first one. Although the character is Geralt of Rivia with predetermined background and abilities, what I experience is my interpretation of Geralt of Rivia, and the game allows me to make choices based on that identity. In a way, even if you can freely create your character at the start of the game, the game itself still dictates a lot about what your character can become - in this sense, starting with a predetermined character is not actually all that different. In fact - to my knowledge - it's actually very common to have pre-written characters in live action roleplaying games. It's not unheard of in tabletop RPGs either.

2. Virtual relatedness 

We already know that I place a lot of emphasis on character writing in games. I know it's a bit silly, but I don't have that high expectations so it kind of works out. However, when we're talking about roleplaying, the importance of other characters in the game rises even further. A lot of choices in BioWare games concern how the player's character interacts with other people. To put it another way, the player's capability to express their projective identity is heavily influenced by other characters in the game. This, in turn, puts a lot of emphasis on a) how well these characters have been written and b) how well interaction with them is presented in the game. Creating interesting characters has been a strong point for BioWare for the longest of times - and Dragon Age is definitely not an exception. Indeed, without a strong illusion of virtual relatedness I would not have bothered with the game for long.

There is one clear sign that Dragon Age succeeds in creating this sense of virtual relatedness: I had very different opinions of companions when I thought about them as myself, as opposed to when I thought about  them as Cath. It is, I think, the best indication that Dragon Age succeeds first and foremost as a roleplaying game. Whereas I find Morrigan's antics amusing, for Cath - whose life depends on her companions - they made Morrigan seem incredibly untrustworthy. While Leliana might have been naive and even childish for me, for Cath she was the most comfortable person to be around. One more: I liked Alistair from the get-go almost, but it took almost the entire game for Cath to be able to truly consider a human male as a friend. All these companions had a huge role in building Cath. Her choices became influenced most by those she held closest, and she became skeptical of anything Morrigan seemed to approve of.

The sense of relatedness is created by making companions (and other NPCs) feel as much like people as is feasible in a digital game. They have their opinions, and they make observations in the environment. They also talk to each other when traveling with the player. Dialogue with them is written well, and it's delivered with very solid voice acting. In short, they make you want to talk to them, even become excited about being able to visit the camp between dungeons. This may in fact have been one reason why I felt the dungeons were so effing long - they kept Cath from having chats with her companions. Indeed, in contrast to camp conversations, the dungeons offered nothing to Cath's personality growth. They were merely obstacles between me and the content I actually liked experiencing. For the record, Mass Effect largely failed to create similar feelings in me because of one important difference.

In Mass Effect, while characters certainly are conceptually sound and interesting, they are portrayed rather horribly. The effort to make them feel like people has been either misguided or simply lacking. Having a conversation with anyone in ME felt a lot like reading an autobiography. That's a fancy way of saying that characters almost exclusively talk about their background. Some banter exists when they are taken with Shepard to missions and space stations, but not enough. I never truly formed any sense of relatedness towards Shepard's companions in ME because they did not feel like people. Even Garrus, probably my favorite character in the series, felt a lot like an audio book. In comparison, characters in Dragon Age feel more real, and thus more relatable - even though they also do talk a lot about their past. The key difference is that they talk about other things more.

Sense of relatedness is not strictly tied to roleplaying. I see it as a facilitator. Relatable characters aid the player in constructing their projective identity by putting the character in social situations - something that for us as social creatures is massively important in defining who we are. That is to say, it's not impossible to roleplay in an empty world. It's simply easier to construct a personality for your character if you can reflect upon that personality through interactions with other people. You know you're there when empathy replaces instrumentality as the player's basis of making choices (this is something I want to expand in another post). Likewise, relatedness is important in all games, not just roleplaying games. This is hardly surprising, as it is just as important when consuming other forms of fiction - be it books or movies. Narratives tend to work rather poorly if it's enacted by unrelatable characters - we need to care.

3. There is no freedom in a group

Although we have now concluded that characters in Dragon Age are quite relatable as people, there is one thing they are not - and that's a group. It did not bother me as much this time around, but it certainly bothered me when I was playing Mass Effects. A lot of dialogue occurs between the protagonist and her companions, but the companions themselves speak very little with each other. There are no group events in the camp - basically all dialogue between companions takes place when they are in the active party. While well written and somewhat frequent, it isn't enough to provide a sense of group. The relations between companions are superficial at best and while their antics are amusing, they don't develop into anything. In this respect, the game is heavily centered on the protagonist. This is actually a very common trait in roleplaying games of Western origin - it's also a cultural thing.

Culture or no, conveying a sense of group in a game with relatively high degree of freedom is far from trivial. By limiting most interactions to one-on-one, the developers have saved a lot of resources. Whereas one-on-one conversations only need to branch based on player choices and a few flags, group conversation branching is likely to get out of hand quickly unless certain limits are placed - and DA is just not placing them, opting for more freedom instead. This is fine, and supports what makes it good as a roleplaying game. It does however also mean that it would be almost impossible to convey a sense of group by having many-to-many interactions. First of all, simply accounting for the fact that there is no telling which characters will actually be present would require a ton of branching. On top of that, they would also have to take into account all kinds of status flags about the protagonist's relationship with each character etc.

I am somewhat curious whether this aspect has been developed further in, say, Inquisition and its war council. Meanwhile, a close comparison can be found from the Persona series, where the sense of group is immensely strong - especially in P4. However, the game is practically linear and while the player character can have different relationships with his companions through social links, these relationships are in no way reflected as branching in group events. It makes the entire group more relatable, but at the cost of player freedom and their ability to roleplay. In this scenario, writers are always guaranteed that certain characters will always be available for the group event, and there are not status flags to worry about. Even player choices during events are there almost exclusively for flavor, so there is no branching whatsoever. This is a common trait in Japanese RPGs and they play out much more like a TV series in this sense.

It may not be impossible to get the best of both worlds but I'd imagine it would be incredibly resource-intensive. With very careful writing, it could be possible to make conversations modular enough that you could simply omit and/or replace single lines without changing the entire discussion at each branch. However, doing this while still attempting to make the dialogue interesting and the characters relatable sounds like a massively tall order. In the future problems like this could be maybe addressed with AI when they learn to produce text in a credible fashion while staying true to the speaker's personality - but again, that sounds far off. Until a game comes along and proves otherwise, I am going to consider the lack of group sense in DA as a cost that comes from giving the player freedom. Therefore, limiting the game to one-on-one interactions has likely been a conscious choice rather than an oversight.


Last time Dragon Age: Origins took a serious bashing from me; this time we have come back to explore its redeeming qualities. What the game lacked in, well, being a game, it took back in being a roleplaying experience. The real story in the game for me was following Cath's growth. Being able to immerse myself in the game world through my projective identity as Cath was far more important than any game mechanical aspects. This, to me, is an appeal that is generally more common in Western roleplaying games but even among them, strong experiences that get close to what tabletop RPGs offer are few and far between. Getting so much out of other aspects of the game most likely made the dungeon parts even more frustrating though. While companions feel like more relatable than they did in Mass Effect, they still do talk a bit too much about their past and a bit too little about other things.

Finally, there is no sense of truly forming a group of people in the game as most interactions happen between the main character and a single companion at a time. It would seem like a necessary amendment that is required to provide the desired degree of freedom to the player.

Because of the gameplay shortcomings I am a bit skeptical about Dragon Age 2, given that it has been generally decreed inferior to Origins. For some reason I still feel bad about skipping a game should I move straight to Inquisition. There was also another consequence after playing DA. Since I really enjoyed the drama and basically everything except dungeons, I decided to finally look into Telltale's storytelling games and also to play other purely story-based games (although, I actually just stopped playing any games at all* for a while after finishing Walking Dead season 1). My next post will probably be about Persona 4 Golden though. I will still write about some story-based games in the future.

* Dota 2 is not a game - it's a lifestyle.

Monday, November 17, 2014

Dota 2 Next Level Meta: Support Bloodseeker 6.82 update

Since Bloodseeker was changed massively in 6.82, I need to revise my earlier post about support Bloodseeker. In fact, he is now so much better as a jungling support. Here's a quick summary of the changes:

His first skill, Blood Rage, is now a combination of the old one and Bloodbath. It's now a buff that can be cast on both enemies and allies, heroes and units. The target deals percentage based bonus damage with everything (i.e. not just right clicks) but also takes bonus damage. When a unit under Blood Rage kills an enemy, they regain 25% of the enemy's HP pool as healing - likewise, if they are killed, the killer gets 25% of their HP pool as healing. Because of this change, Bloodseeker is now a much stronger jungler, as the healing is much higher now on level 1, and it also grants bonus damage, allowing him to clear camps quicker. As long as you make sure to have the buff enabled whenever you deal the killing blow, jungling is a piece of cake. Chokepoint jungling is still recommended. If facing multiple enemies, it's best to cast this on your target instead of yourself, as the other creeps would otherwise also deal bonus damage to you.

His second skill, Blood Rite, is mostly new. The silence from old Blood Rage is now inflicted by this skill, but it's shorter. However, it also has a huge area of effect and deals pure damage, but with a significant cast delay. This skill needs some setup, obviously, but it's a strong nuke and it can be used to zone out enemies efficiently even should they be able to avoid it. Naturally you can set it up using Rupture but it's even better if your team can provide some setup. His third skill, Thirst, is the same in principle, but now provides a gradual bonus instead of using a threshold, which means every bit of missing health now counts. The vision threshold has been reduced to 30%. Rupture now lasts practically forever (ok, 12 seconds), but it has no initial damage. It's still more or less the same skill. Town Portal scrolls are still the bane of his existence.

So, how does all this change things?

First of all, you only need one skill point for jungling, since level 1 Blood Rage will suffice easily. I would leave it at one point until the other skills are maxed, as it scales pretty badly, and percentage based bonuses are not too hot in the early game. Since you can cast it on a target, you can allow all of your team to make use of the bonus damage - just remember that the target will also be dealing bonus damage to all of you. You can also use it on your carry like the old version. Instead of silencing them, you'll put them in higher danger however - so don't just throw it out when your Sniper is getting whacked by the enemy's Phantom Assassin. As for what to max first, I'm divided between Blood Rite and Thirst. As a support it's generally better to max your nukes, but Blood Rite has pretty bad scaling. If you need the longer silence and have some setup to land it, go for it. However, even a level 1 Blood Rite has a lot of potential as a zoning tool, and Thirst has very good scaling.

As for items, Mekansm is no longer viable on Bloodseeker because of the increased mana cost. This is somewhat unfortunate, as you could have built a pretty fast mek with the changes to Blood Rage. The only way to sustain a mek and your other skills would  be to build Arcane Boots, but it doesn't really sound all that hot. Instead, you might want to pick up a Force Staff - for all the reasons mentioned earlier - or as a new suggestion: Eul's. You can use it to cancel those pesky TPs, run even faster and setup your Blood Rite on a target (heroes with blinks, leaps and similar abilities can still get out). Diffusal Blade or Rod of Atos can be useful too - Diffusal is still probably better, but Atos does make you tankier. Blade Mail is still strong. Another note about Blade Mail: if you make a habit of casting Rupture on Blade Mailed targets, getting a Black King Bar might be smart. Or, you can also Eul's yourself when they start to run.

As for the game plan, it's mostly intact. Go to the jungle, but don't just afk there. Once you have a level in each of your abilities, you can easily gank lanes (remember to carry those TPs!), or even save a teammate with a zoning Blood Rite. You might even want to leave it at 1 and max Thirst, and then just play mind games with it as enemies are still very likely to try and avoid it. All in all, I think support Bloodseeker is now fully legit if you just jungle and roam, and build some utility items.