Extra Credits: Balancing for Skill

[SinbadEV]

The latest Extra Credits episode seems quite relevant to what I have peripherally observed of the balancing and n00b retention problems people tend to complain about 'round here.

http://penny-arcade.com/patv/episode/ba ... -for-skill

[gajop]

saw this, seems stupid
one just doesn't ruin a game balance to cater to noobs
if you want noob retention you create ELO-like rating systems and have people playing (either co-op or competitively) with other people at that same skill rating
creating any kind of game mechanic which will dominate a lower skill bracket is just bad design, seen in many games, often in ways of stealth/high dmg burst/uknown mechanics that devastate low skilled players and are nearly always useless as you skill up

[smoth]

saw this, seems stupid
one just doesn't ruin a game balance to cater to noobs

Way to see it from the extreme other side.

watched this already sinbad... all I can say is HA:
BA DSD why it works, how it is good and you getting over it.

[Jazcash]

The latest Extra Credits episode seems quite relevant to what I have peripherally observed of the balancing and n00b retention problems people tend to complain about 'round here.

http://penny-arcade.com/patv/episode/ba ... -for-skill

Nice video!

[gajop]

Way to see it from the extreme other side.

From what I got from the video, it was all about how noobs/casuals would see the game:
part 1 was about how they need a easy-to-use mechanic that rewards immensely initially
part 2 was about how that easy-to-use mechanic shouldn't be too good otherwise those noobs would give up when something comes up/gets changed and they can't use it anymore

[smoth]

Way to see it from the extreme other side.

From what I got from the video, it was all about how noobs/casuals would see the game:
part 1 was about how they need a easy-to-use mechanic that rewards immensely initially
part 2 was about how that easy-to-use mechanic shouldn't be too good otherwise those noobs would give up when something comes up/gets changed and they can't use it anymore

It is about how you have to balance the game so that the noobtube while good becomes less valuable forcing them to eventually move to something new. OTHERWISE they end up realizing all to late that the mechanic they relied on didn't teach them the skills they needed at higher levels of play.

[SwiftSpear]

I've been thinking a lot about this. I made a couple graphs.

An ideal skill curve looks like this:

Some wiggliness is ok, probably a good thing actually, but the graphing program I used doesn't like that type of thing very much. Irregardless here are the main points

- It doesn't start at zero, the game needs to hand hold you in the beginning so you don't feel like a retard, the skill floor to start playing the game is as low as possible
- the early progression gives big power gains, you feel like getting better at the game is a good thing to do
- There are no massive discrepancies, no cliffs where either practicing doesn't help you any more, or you suddenly open up a can of hellishness by mastering a certain skill
- Most importantly, it doesn't "plateau" by decreasing the amount of power you gain by gaining more skill. The ceiling never levels.
- Your first 300 some odd hours of play give you a 2-7 power jump, your second give you a 7-9 power jump, every 300 hours after that it's less than a full point, but no matter how much work you put in you always see some pay back

The reason why gaming is in a mess, because most "skill based games" have either made this mistake:

or they have made the mistake of starting the power graph at the zero point.

And most casual games do this:

- just don't ever bother putting more than whatever level X amount of hours is into this game at which you hit the plateau.

I can't think of a game that I think has done this well outside of maybe Othello or checkers. Video games all seem to hate being comprehensive experiences.

[Johannes]

The reason why gaming is in a mess, because most "skill based games" have either made this mistake:

That's not really the full story. Sure, a lot of games have techniques like that which are really useful but you won't be using them at all at first, be it bunnyhopping in FPS, backdash canceling at Tekken, muta stacking in SCBW (not saying whether these individual mechanics are necessarily good or not). But it's not like you can't improve endlessly in other areas of the game even if you'd forgo these things. And it's not an on/off switch to master them either, you can still near endlessly improve on how well and reliably you can do these things.

Or just using hotkeys and ctrl groups in an RTS - everyone should do this yet they probably don't use them right from the start. You can still improve without them, though. Like you can do ok without widgets in Spring games if you're so inclined.

It's just natural that the scope of things you should focus on and do, increases as you get better.

[gajop]

I could only agree with that if the Skill axis is on log scale.. otherwise you would want power to go exponentially up. The real reason why, is that there's a limit of how much humans can be skilled, and going from skill level (whatever that is) 1 to 2 takes some x effort, while going from skill 2 to 3 will often take some n * x, n > 1.

[PicassoCT]

Demanding skilll improvement is important, i agree on that, but the experience as a whole is dependent on the hollywoodcurve.. means tension (more skill), pause (relax) and you can design for stuff that is paradox. Fear if done right can freeze the reflexes of the most talented player, thus design for emos-ocean ripples ..

[SwiftSpear]

Demanding skilll improvement is important, i agree on that, but the experience as a whole is dependent on the hollywoodcurve.. means tension (more skill), pause (relax) and you can design for stuff that is paradox. Fear if done right can freeze the reflexes of the most talented player, thus design for emos-ocean ripples ..

I agree with you, but we're talking about something different. Pacing and 'skill vs power' ratios are separate considerations. My graph is more representative of hundreds or thousands of games as opposed to the demands of one game. Skill curve is also less important in single player games... I'm focusing pretty heavily on the competitive multiplayer component. (not competitive as in playing MLG, just as in there are other people I am trying to win against)

You're right, that's a mistake a lot of games make, they try to "remove" all the boring stuff, and you just end up with an experience you find stressful and you never have time to relax. If you stick through it the experience is more of a grind than it is an accomplishment of tasks that you are happy about and relieved to finish. And the great thing is that you also want to preserve that pacing no matter how skilled the player is. So if a really skilled player is playing on hard, they still find it difficult to just shoot one guy and skip by the main conflict scene, (although there ALSO reaches a point where this is forced and annoying when speedrunners get their hands on your game, and ideally you cater to them as well , although they're usually happy as long as you don't artificially force them to slow down)

[SwiftSpear]

I could only agree with that if the Skill axis is on log scale.. otherwise you would want power to go exponentially up. The real reason why, is that there's a limit of how much humans can be skilled, and going from skill level (whatever that is) 1 to 2 takes some x effort, while going from skill 2 to 3 will often take some n * x, n > 1.

I think it depends how you measure skill. If you measure skill by some metric of density of valuable inputs into the game (a glorified APM of some type), then I agree with you. If you measure skill more along the lines of "this is the power the average player gains by investing 100 hours of practice into the game" then my system is more on point with what I'm intending.

But yeah, it's interesting, the difference between 15 APM and 200 AP for a game like starcraft for example is much more realistic and acheivable than the difference between 200 APM and 385 APM. Many players practice their whole lives and still don't reach that level.

[gajop]

If you measure skill more along the lines of "this is the power the average player gains by investing 100 hours of practice into the game" then my system is more on point with what I'm intending.

Then we are in agreement of that graph, just with disagreeing definitions on what skill is. I still wouldn't define skill as time invested to obtain it (since that's usually not the same for each person), but would rather try to define it as the player's effective capability in the game: APM is a one such metric.

A large part of a player's skill in a new game is his past gaming experience and IQ/hand-eye coordination/dexterity. Time invested to play a certain game is usually not very important for old gamers.

Example: yesterday I played Rayman Origins on the XBox, without ever playing Rayman 1 or Origins, without owning XBox or any other console, etc. I played my last platformers when I was a kid (lion king/aladin/mario/prince of persia time).
I was still one of the better players, since other people had very little to no gaming experience, even though they had played (in the past few days) that game for over 6h.

[smoth]

Then we are in agreement of that graph, just with disagreeing definitions on what skill is. I still wouldn't define skill as time invested to obtain it (since that's usually not the same for each person), but would rather try to define it as the player's effective capability in the game: APM is a one such metric.

A large part of a player's skill in a new game is his past gaming experience and IQ/hand-eye coordination/dexterity. Time invested to play a certain game is usually not very important for old gamers.

I would put the iq/cordination in talent. Skill is something learned IMO.

[SwiftSpear]

Then we are in agreement of that graph, just with disagreeing definitions on what skill is. I still wouldn't define skill as time invested to obtain it (since that's usually not the same for each person), but would rather try to define it as the player's effective capability in the game: APM is a one such metric.

A large part of a player's skill in a new game is his past gaming experience and IQ/hand-eye coordination/dexterity. Time invested to play a certain game is usually not very important for old gamers.

I would put the iq/cordination in talent. Skill is something learned IMO.

Agreed for the most part... Strictly speaking, talent seems to be more an ethereal ability to pick up something easily, where as IQ or coordination aren't really specifically talent (although they may make you appear more talented). It's sort of an innate biological leverage.

However, generally when players talk about "skill" in a gaming environment, they're talking about a combination of talent, natural advantage, and actual skill. Professional Starcraft 2 players will attribute their success to "skill" however, it's hard to argue that they don't have an inhuman ability to quickly process RTS style problems intellectually, as well as mechanically move their hands accurately and rapidly far superior to what most people could ever accomplish even with training. They will admit that as well, that's just not what they attribute as the majority of their success.

[smoth]

at that level, yes, they have better muscle memory but not many of the people around here are at that level. Most of the people at that level are not around much any more or seldom participate in these sort of threads.

That is why I find them so frustrating. People talk about spring like they are at the height of skill but in reality most of the are just mid level players. there is a lot of pushing for really high level stuff but what I find they really want are obscure mechanics that give them an unfair advantage because they either found out or lucked up and were clued in to these mechanics.

so they seem to push heavily for low skill obscure measures it is like security through obscurity. If say ba was a bigger more competitive game there would be a real wiki discussing the dsd meta and obscure mechanics naming them.

[Johannes]

However, generally when players talk about "skill" in a gaming environment, they're talking about a combination of talent, natural advantage, and actual skill. Professional Starcraft 2 players will attribute their success to "skill" however, it's hard to argue that they don't have an inhuman ability to quickly process RTS style problems intellectually, as well as mechanically move their hands accurately and rapidly far superior to what most people could ever accomplish even with training. They will admit that as well, that's just not what they attribute as the majority of their success.

Inhuman, really? There's nothing inhuman at merely being better than most at something. I know it's just a quick choice of words but...


Though what I think is the most important thing at the highest level, is to be able to practice for long periods and not get mentally fatigued by it, but constantly being able to figure out new things.
And hand dexterity is overvalued a lot. If almost anyone can be taught, for example, to play the piano decently given enough practice, getting an RTS interface to your muscle memory isn't impossible either. The hard part is knowing what to do without wasting time (both on micro and macro level, and not meaning the usual RTS terms here), executing that is quite easy by comparison. Muscle memory comes with time, for anyone.

[Silentwings]

If almost anyone can be taught, for example, to play the piano decently...

They can't! But I have no idea what kind of dexterity pro gamers have.

[SwiftSpear]

However, generally when players talk about "skill" in a gaming environment, they're talking about a combination of talent, natural advantage, and actual skill. Professional Starcraft 2 players will attribute their success to "skill" however, it's hard to argue that they don't have an inhuman ability to quickly process RTS style problems intellectually, as well as mechanically move their hands accurately and rapidly far superior to what most people could ever accomplish even with training. They will admit that as well, that's just not what they attribute as the majority of their success.

Inhuman, really? There's nothing inhuman at merely being better than most at something. I know it's just a quick choice of words but...


Though what I think is the most important thing at the highest level, is to be able to practice for long periods and not get mentally fatigued by it, but constantly being able to figure out new things.
And hand dexterity is overvalued a lot. If almost anyone can be taught, for example, to play the piano decently given enough practice, getting an RTS interface to your muscle memory isn't impossible either. The hard part is knowing what to do without wasting time (both on micro and macro level, and not meaning the usual RTS terms here), executing that is quite easy by comparison. Muscle memory comes with time, for anyone.

Playing the piano decently is not the same thing as playing a professional RTS. What pro players eventually master is much more akin to what elite level pianists perform, not ANYONE can be trained to play at the level demanded of the London Symphony Orchestra, for example. Most people can achieve a lot more through muscle memory than they would expect, but there is a line between truly elite players and the majority, and it doesn't entirely come down to hard work.

It's just like any other physical sport. Mike Tyson could never be an Olympic gold medal swimmer, and Micheal Phelps could never be an Olympic gold medal gymnast. People are good at, and talented at, things their minds and bodies tend to trend towards for some reason. Hard work can do A LOT but it can't completely invalidate genetics.

[SwiftSpear]

at that level, yes, they have better muscle memory but not many of the people around here are at that level. Most of the people at that level are not around much any more or seldom participate in these sort of threads.

That is why I find them so frustrating. People talk about spring like they are at the height of skill but in reality most of the are just mid level players. there is a lot of pushing for really high level stuff but what I find they really want are obscure mechanics that give them an unfair advantage because they either found out or lucked up and were clued in to these mechanics.

so they seem to push heavily for low skill obscure measures it is like security through obscurity. If say ba was a bigger more competitive game there would be a real wiki discussing the dsd meta and obscure mechanics naming them.

I think often people confuse obscure mechanics for high skill mechanics. Things like adding spell effects to RTS units. It's not really obscure what the spells do, but mastering their usage requires more clicks and more accurate selection on the map. You can't just attack move those units anymore and expect them to perform at maximum efficiency. Or dealing with fast units that have a slight range advantage over a slower opposing unit. It's not really at all obscure that they CAN beat that opposing unit, but if you intend to use said fast unit against that opposing unit, you need to control that unit well enough that it can exploit the advantage it has.

There are obscure things as well, like mutalisk bunching up micro from Brood War. It served a role once it was discovered, and it began to catch on over time, but it was CLEARLY an obscure mechanic, in fact, not one that was even intended. I think, as a game developer, love and utilize the former two to the best of your ability, and avoid and restrict the latter one as much as fair and possible.

This is also the same reason why bunnyhopping is bad in games. I LOVE bunnyhopping from the halflife and quake 2 era. It's SO much fun once you master it, it's a skill mechanic that is beautiful in that it takes some time to learn the basics, but no matter how hard you practice you will never have it "perfect" there's always more improvement to make... But it's incomprehensible to organically figure out. It's the exact definition of an "obscure" mechanic. It has to do with understanding how air control mechanics work in those games, and even then, you need to not just understand it, but also practice a mechanical skill that really has no real parallel to anything done in real life in any way.

What it shows us is the beauty of a design that allows for a high skill ceiling, while also showing us the dangers of sacrificing our game to an obscure skill based system.

RTS in general have a higher skill floor. None the less, I think it's very bad design to deal with this by building a skill ceiling lowerer into your game, as if the skill floor and skill ceiling are the same thing. Granted in many cases they can scale together... but one of the higher priorities of game design for multiplayer games should always be pushing the skill floor and the skill ceiling as far apart as possible.