SC2Mapster

If so could you please provide some feedback here:

http://finalascension.org/viewtopic.php?f=30&t=103

I really love math because you can create really complex formulas out of really simple things :p

It looks interesting anyway, going to read it.

Btw, wtf the "Do not copy" on everypage? Seriously I want to give remarks & show typos but I can't copy & paste things ...

Edit:

Okay just read the whole article. It is interesting.

Proof? Your theory is great but you could say full bullshit it would be the same. There is no theorem proving, no verification of any kind on examples. Have you tried to parse replays and apply your models on real data and see if it fits? Where are the benchmarks & such things :)

References You are talking several times about Game Theory but you are not using it anywhere. You have references but they don't appear anywhere either.

Mathematics You say it's a mathematic paper however the only concepts you are using are only basic operators (*, +, /, sigma). You are using regression at the beginning but just defining the concept and it isn't used after.

I really think this kind of article is needed however I think it requires more investigation. You defined many concepts but you don't prove them (nor show they work on examples) and start from scratch without reusing any other paper at all.

I hope you are going to continue it because I really think it has potential.

Edit2:

I'm not sure that you should put this as a reference ... http://mlewi.com/content/projects/INFOGD.4/Documentation/Game%20Design%20-%20Design%20Document%202008-2009.pdf

The CNN article is from a blog and ... there is no argument given as for how the game is unbalanced. How would it be a valid reference !? http://scitech.blogs.cnn.com/2010/03/19/geek-out-starcraft-ii-wings-of-liberty/?hpt=Mid

Quote from vjeux: Go

I really love math because you can create really complex formulas out of really simple things :p

It looks interesting anyway, going to read it.

Btw, wtf the "Do not copy" on everypage? Seriously I want to give remarks show typos but I can't copy paste things ... Edit:

Okay just read the whole article. It is interesting.

Proof? Your theory is great but you could say full bullshit it would be the same. There is no theorem proving, no verification of any kind on examples. Have you tried to parse replays and apply your models on real data and see if it fits? Where are the benchmarks such things :)

References You are talking several times about Game Theory but you are not using it anywhere. You have references but they don't appear anywhere either.

Mathematics You say it's a mathematic paper however the only concepts you are using are only basic operators (*, +, /, sigma). You are using regression at the beginning but just defining the concept and it isn't used after.

I really think this kind of article is needed however I think it requires more investigation. You defined many concepts but you don't prove them (nor show they work on examples) and start from scratch without reusing any other paper at all.

I hope you are going to continue it because I really think it has potential. Edit2:

I'm not sure that you should put this as a reference ... http://mlewi.com/content/projects/INFOGD.4/Documentation/Game%20Design%20-%20Design%20Document%202008-2009.pdf

The CNN article is from a blog and ... there is no argument given as for how the game is unbalanced. How would it be a valid reference !? http://scitech.blogs.cnn.com/2010/03/19/geek-out-starcraft-ii-wings-of-liberty/?hpt=Mid

That is perhaps the smartest response I've gotten so far. You bring really interesting points.

@Copy&Paste, Sorry, its an old habit from college. If you would like I could send you a .doc

@Proof: That is a valid concern. However, it is stated it is a theoretical framework. For the regression I cannot obtain the data unless I apply direct observation to each reply. For instance BW chart includes repetitive actions as a part of micro and I state that in most situations this is a bias.

The C-Matrix has a tier1 example of SC

The Resource counter had an example with SC2 regular units.

@Game Theory reference: I did mention it once, regarding the 2v2v2 example.

@Mathematics: And those operators are perfectly valid to label it as such. The regression is 1/2 of the paper.

@CNN. You are correct. They don't provide any argument! That is the point of this paragraph: There is no existing methodology to formally evaluate if a game is balanced or not. I have spent days looking for RTS balance design articles and found nothing useful that is why I'm starting from scratch. It is my very humble opinion that in a new field of study such as this one, papers will dull your mind and destroy the potential for original creativity.

"Recently an article on CNN (perhaps not the best source for a game review) stated that StarCraft 2 had an “unbalanced game play, based on which race you play”. The article failed to mention and justify how and why the game was unbalanced. Still they provided an opinion without justifying their arguments. Quite simply there isn’t a formal methodology or framework to determine how balanced a game is other than very simple statistics and subjective opinions. This was part of my motivation into writing this paper."

In any case, I will take these things into consideration and update this. Thanks a lot for your input, vjeux. Any more suggestions please keep them coming! :)

What is your definition of a "balanced game"? (sorry if I missed it :<)

@RandomNoExit: Go

I didn't explicitly define it. I'll add it on the next revision. To answer your question:

A balanced game is defined as a game that does not have any dominant strategies. In the case of an RTS it would revolve around the idea that there is no dominant race, unit or spell. Dominance is a game theory concept. A pure strategy is dominated if there is another strategy where the utility of a player is greater for all the strategy profiles of the other players. In an RTS this means there is no strategy (rushing, teching, mass-producing a particular unit) that will always win the game given all the possible strategy profiles of the opponents.

Thanks for the definition.

The problem though that I have is that I miss something in the definition that is related to a map. Certain strategies work "better" on some maps while others don't make sense at all (if there is a map with no vespene gas, then a "carrier strategy" wouldn't really make sense.) I'm pretty sure there are many mirror maps possible that do have dominant strategies. (example island map with 2 islands where they are so small that terrans can't build a lot of supply depots and protoss can't build a lot of pylons, but just big enough for zerg to get mutalisks ... I know I'm evil.)

I would also add in this definition that every pure strategy should lose against at least 1 strategy per race.

Another question I'm asking is what properties do the players have and what effect does it have on the strategies (for example strategies which require superman like reflexes, are they valid?) and what effect it has on the definition of game balance? (Note also that it might be possible that an early 5 marine rush with perfect micro can kill 3 zealots: the 3 "chased" marines run around and the 2 other marines shoot.)

Quote from RandomNoExit: Go

Thanks for the definition.

The problem though that I have is that I miss something in the definition that is related to a map. Certain strategies work "better" on some maps while others don't make sense at all (if there is a map with no vespene gas, then a "carrier strategy" wouldn't really make sense.) I'm pretty sure there are many mirror maps possible that do have dominant strategies. (example island map with 2 islands where they are so small that terrans can't build a lot of supply depots and protoss can't build a lot of pylons, but just big enough for zerg to get mutalisks ... I know I'm evil.)

Thanks for your response.

You mentioned islands, and in my paper I consider the map-type as a different balance dimension that can change the strategy profiles for players. We probably know you won't likely be seeing that map on ladder :P

In my original definition just add a "ceteris paribus".

Quote from RandomNoExit: Go

I would also add in this definition that every pure strategy should lose against at least 1 strategy per race.

Yes, that could be a valid corollary I believe.

Quote from RandomNoExit: Go

Another question I'm asking is what properties do the players have and what effect does it have on the strategies (for example strategies which require superman like reflexes, are they valid?) and what effect it has on the definition of game balance? (Note also that it might be possible that an early 5 marine rush with perfect micro can kill 3 zealots: the 3 "chased" marines run around and the 2 other marines shoot.)

Skills is obviously a valid and relevant variable. You bring a valid point but the question is, how would you measure that particular type of skill?.

Quote from Genopath: Go

Skills is obviously a valid and relevant variable. You bring a valid point but the question is, how would you measure that particular type of skill?.

There's this little thing called APM (actions per minute) it separates the pros from the newbs. The other thing i would check is build orders, and specifically average unspent minerals/vespene. You could even check unit, structure, and resource scores.

Edit: I'm taking a course on Game Theory this summer, god i hope they mention starcraft in it.

@Sixen: Go

That is correct but it depends a lot more on the sample. If your sample contains bronze, copper, silver, gold and platinum games. You might see APM as a significant variable affecting the outcome of a game.

If your sample contains only platinum games I guarantee you APM won't be a significant variable because at this point having an APM above 300 or 400 doesn't make much difference in the outcome of the battle. Sure it can mean someone's reflexes are faster but it also can mean player's like to spam a lot.