Devs called I24 "Fix Everything" in Coffee Talk... Lets hope that doesn't include nerfs

Off Topic: So do you think big companies like Blizzard hire mathematicians then? Say for World of Warcraft? Or would you say that game is on the same complexity level of COH?

I actually worked on a model for the stacking resistances that produced consistent results, but it occurs to me the forum I posted it on is currently in "archive" mode and I can't access it. Basically, I chose to represent "resistance x -health" as a 2D graph, with the area that's "below" curve of the health-dependent resistance constituting average resistance across the interval. It's easy enough to calculate with a straight-line curve, but this model also gives you the ability to calculate it for an actual curve via definite integral. That's basically what a definite integral itself models.

That aside, what I'm saying is I don't think balancing a game around math in the first place is a good idea. You just put the onus on planning and winning fights before they even start. Sure, you can make it complex and obscure enough to where most people won't get it, but at that point you're just hiding the right way from your players while still HAVING a right way built into your game. It's why I've been more and more disappointed with stat-driven RPGs of late.

Again, every time I play an actual action game where new abilities actually give me the ability to do new things as opposed to just modify some numbers, this is reaffirmed to me. A new ability with a different function changes how I play in an action game. In an RPG, it typically just changes my stats for doing more or less what I was before, just by clicking different-looking icons. Look at something like League of Legends. For as much as I don't like that game, there really is a difference in how all of its characters play, not just what stats they bring to the table. And that's not just the math talking.

Originally Posted by Arcanaville Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

I'm not a WoW expert, but based on what I do know about WoW I would say that fundamentally speaking WoW is a much simpler game to analyze. There's more "stuff" but the stuff itself is not difficult to figure out what it does precisely.

CoH combat is actually very complex to analyze, relative to practically all other MMOs. And that's because most MMOs are built on models that make them easy to analyze so the *developers* can predict what will happen. The original CoH devs got that one horribly, horribly wrong in this game. They created a game that exceeded their ability to analyze. Eight years later we're still developing the analysis tools to analyze this game.

And then the next step is to have the game analyze itself...and then we've reached the Singularity.

OMG CoH is Skynet!


.

This is just one of those things where I've come to realize game designers aren't students of their own field. When he says he doesn't balance the game, he's simply introduced a different balancing system into the game: one where powers and abilities degrade in relative ability with progression, so there's an automatic negative reinforcement that the more powerful an ability is in terms of its ability to make you level, the less amount of time you can use it for - because it will outlevel itself.

Around these parts we system engineers call that "intrinsic reward rate-based balance."

And if he really doesn't balance the system, why is he concerned about ironing out the "crazy peaks?" Because he does care about balance, he just wants to do it the long way.


Also, the really different way he thinks he's doing it now is how Cryptic and now Paragon Studios has done it for about ten years, initially because it was the only way they knew how, and now because it would be too complicated to do it any other way. Arbiter Hawk seems to be the most number crunching dev we've had since Weirdbeard, but he doesn't use that number crunching to "balance" the game in the conventional sense, but to generate the results he's aiming for, which aren't necessarily what traditional balance would dictate. Its still very much balance by iteration.

Well what really strikes me about his approach is that he actually usually ends up with a finished game. For a small publisher, just getting the game out the door on a reasonable budget and having it not totally suck is a major feat.

There are so many small things that can wreck an RPG (action or not). I don't blame him for not running complicated models beyond basic assumptions about stat distribution, because in the end it won't matter if you find out one class can consistently use skill to dodge the monsters completely, because of the distribution of the monsters or gear, not because of the mathematical model underpinning the game. This specific scenario has played out in most games in one form or another.

The time "lost" up front would almost certainly be paid back with interest very quickly, since less guess, test, guess, test would be necessary. Oh, you would still need to test, but if it was done right, there should be less iterations needed and rework is the bane of meeting schedules, usually much worse than process development.

I have had bad experiences with process development, but that doesn't mean it should be avoided or minimized.

I think the problem is people oversimplify the issue with RPGs specifically. I am not saying you shouldn't design with some idea of how numerical systems will operate. The City of Heroes defense calculations are a testament to that.

But I also think there is a widespread idea that even before a product launches you should be able to tell from a spreadsheet exactly how combat mechanics will play out. There is really only one way for that to happen in reality: your system has to be be highly predictable. Unfortunately, sustained fun and high predictability do not frequently get along. There are plenty of very well balanced Korean grind MMOs, and not a whole lot else.

To put this in CoX-explicit terms, anyone who says they can predict combat outcomes from a system where players are allowed to Fly, Super Speed and Teleport without actually interacting with those systems in a typical game environment is simply mistaken. It's impossible.

We also have to remember that we are not talking about a project a single individual would be creating. Game balance is a product that is a vision of a team, often driven by a Project Manager who only has so much time to review decisions. Over the long term, in an MMO in particular, it's not possible to maintain complete oversight of every system. The only way around this, once again, is to use very predictable gear/monsters/potions/etc.

A sequel to one of Schaefer's games that his team didn't participate in was recently released (I can't name the game by name I suspect) and did exactly this. It's a phenomenal flop--a gear grind game with boring gear, and where because of character movement speed and how that interacts with powers a subset of players was able to sidestep many of the excessively tuned abilities.

A final point to remember is that most games (and MMOs in particular) are broken at their most vulnerable spot. For RPGs this is usually the "farming spot." All that work you put into balance doesn't mean much if players can find an exploit in one place they can farm. Exploitation of this kind isn't a feature of combat mechanics per se, it's usually more a feature of level design, critter distribution, and AI.

[EDIT: I just re-read this and it sounds a little preachy. Not trying to lecture, I just have a lot to say. My apologies if it sounds like I'm chastising.]

The world worked fine in the days of Roman numerals. It just works far better with Arabic arithmetic.

I don't doubt that its possible to create a good game without a strong mathematical foundation. I question why its presented as an either/or proposition more times than not. You can make a good game in BASIC. But I would be astonished if the game development community, and only the game development community, constantly recruited programmers who only knew BASIC just because its possible to make a good game in BASIC. Its possible. It also permanently anchors you in the dark ages of what's possible.

One person making a game in BASIC is a curiosity. The whole industry doing it is mass insanity.

Well again I don't think he's saying (and I'm definitely not) that you don't want good calculations underpinning your design. The City of Heroes Defense system is a mess because of that sort of thing. Although what's interesting about that is the City of Heroes Defense system (ignoring Resists and Regen and so on for the moment) is essentially identical to the one used by Dungeons and Dragons for the past 25 years (1d20, 1 always misses, 20 always hits).

What's different about CoX versus D&D is the enemies.

You're mistaken.

There are many parts to the game that produce visually different results. That produce mechanically different results. That have different functional gameplay.

But that's irrelevant. Eric Schaefer does understand this fundamental point about game design:

I know what Power Blast will do in the game. And I know there's no way to teleport yourself a better version of Power Blast. I know if you run away from critters you'll take less damage. I don't have to care how many possible ways there are to run away. I only need to know that while you run away they can't really shoot at you effectively, and you can't shoot back effectively. I also don't have to be particularly accurate with my guestimates on how much loss in effectiveness that situation has.

This game, like most games, is ultimately balanced around progressional rewards. XP, in other words, and secondarily influence and drops. Its possible to determine optimal rates of earning XP under different conditions that the game balances for. Its possible to declare situations out of bounds on the high side that are tolerated but not balanced for. And its also possible to declare situations out of bounds on the low side that because they are on the low side, we don't care about them at all. If you want to hover snipe your way to level 50, since that's ten times slower than normal soloing I don't care if the critters can't shoot back.

I don't have to analyze how every single piece of the game can possibly function, any more than I have to track every air molecule in the sky to know its going to rain. I especially don't need to track every droplet of water in the sky to know that if I go outside I'll get wet.


Its not particularly easy. Its not spectacularly hard. Its just not done. I'm sure gaming history is full of people who have made the claim that it was possible and then failed miserably to actually be able to do it. That impresses me not even a little bit.

Actually, its two things. One: human GMs that can bend the rules. Two: the rate of combat.

Setting aside regeneration is actually something you can't validly do when comparing the systems. How fast do D&D characters regenerate? That's actually the most important question you can ask about a combat system: what is sustainable? In most pnp games sustainability is extremely low. In CoH its far higher normally, and can be made even higher than that. That one difference changes everything.

What I was actually referring to is the absence of a soft cap. It doesn't exist because what amounts to enemy ToHit/Accuracy is highly variable. You can generally hit a point where you can expect most enemies to miss, but enemies with abilities that boost Accuracy and Defense are extremely common. Enemies with extra defense are also much more common where in CoX they are nearly absent or at least highly avoidable.

Some smearing of the soft cap does happen in CoX but the context is key. What's the probablity of finding a relevant group in CoX that cannot reduce your defense or increase its accuracy? What if in our theoretical MMO that group is also only really encountered on a far off island? What if that island is only really easily accessible to one class (maybe not even the one known for fighting)?

The mechanics need to basically work but the mechanics are not the game without any thought about level design, character mobility, or AI. PVP, where the AI is basically tuned to extremely high levels (according to some ) tends to reveal how Stealth and Mobility can impact gameplay in the extreme, and sometimes in ways that require you to modify the values predicted by a model that cannot account for these elements.

The reason why the system of +5 sword of thwacking vs +4 armor of unthwacking works is actually because GMs consciously ensure that the range of combat makes sense during play. You don't have players with +3 power scooters going up against enemies with +30 hand grenades. Which we had here. A human being wouldn't allow that to continue.

I'm not so sure. The Harm spell required a touch attack and drained you to 1 HP in one cast. If you dodged it I think you still took half damage. Poor dragons and mind flayers never knew what hit them. I miss being that overpowered.

(This is to say nothing of the various instant death spells that peppered the earliest versions of the game and the arduous process of getting a resurrection, if you got one at all.)

But could you clear a map with it for full XP?

No, and unsurprisingly that ability appeared on a class that an edition prior had been considered underpowered... sound like any arguments we hear around here sometimes?

What's craziest about it is that spell actually made it into the video game versions. So my Cleric would run up to a dragon with 4 or 5 copies of it loaded and just spam it until the boss instantly died.

In many video game representations of that ruleset, it's also possible for spell casters to simply outrun melees and never get hit or for melees to run away from spells while they are still "forming". In PVP in this COX actually this also happens; the "AI" (i.e. the other player) is generally smart enough to run out of "patch" powers of any kind if they can, which dramatically affects those powers and their value in the game.

This also strikes me as a balancing system centred around negative reinforcement, and negative reinforcement as the sole motivating factor is why Diablo 2 is one of my most hated games of all times. "You feel strong, eh? That's OK, in a little bit you won't be." Having the game keep yanking my toys out of my hands is the fastest way to pull me out of the experience, and when I get pulled out of the experience I tend to stop thinking like a player and start thinking like a customer. It's less about "Is this game fun?" and more "Is this past time worth the time and money I'm investing in it?" I have yet to find an instance where I've asked the latter question and answered "Yes." with the possible exception of this very game.

Originally Posted by Arcanaville Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.

I don't think the question is whether a game should be based in maths or not so much as it's the question of how big of a role do we want the precise math to play. For instance, games built heavily around skill can, to a large extent, be balanced by gut feeling and by actually playing them, then handing them over to people to play and going from there. Sure, the numbers in these games matter, but they matter in the ways people use them, thus what the game is balanced around is less how then umbers work out and more how people tend to play it.

Very simply put, you can have a "thing" that's very overpowered and still have it balanced if it's enough of a pain in the *** that most people couldn't do it or wouldn't want to bother. Basically, the "rocket jump" of a lot of 90s FPS games. It could be very useful when done by very good players, but hell if I ever found a use for it.

Originally Posted by Arcanaville Samuel_Tow is the only poster that makes me want to punch him in the head more often when I'm agreeing with him than when I'm disagreeing with him.