TvZ/ZvP/PvT stats on maps over time

On March 28 2011 17:32 Catch]22 wrote:
Wow that is pretty cool, would be neat to see one for Destination too.

On March 28 2011 17:38 milikan wrote:
are these stats all collected using tlpd?

i guess you have a program that yanks out the stats per map and graphs it?

On March 28 2011 17:42 Catch]22 wrote:
Whoa, knew Desti was bad for T, not that bad :S neat!

Thank God Fighting Spirit became the new Python, could you do one for python? Would be neat to compare them too.

On March 28 2011 17:47 milikan wrote:
suggestion for next "program": compare the distance of the values for the winrates from 50% for each map as a way to measure balance, then find the most balanced map since 2006. ^^

could graph balance over time, but i think that would be random. map makers come up with new concepts, and it seems like they care more about creativity than balance. which i kinda like!

[edit] in the sidebar, i saw the title "race relations..." and i thought this was a post on human race relations. thats how it caught my attention.

On March 28 2011 17:48 Crunchums wrote:
wuthering heights? battle royale? raid assault?

sadly i don't think there are very many maps with a large enough sample size for those graphs to mean much, but they are still fun to look at

hopefully FS will come back during PL playoffs ( <3 FS )

On March 28 2011 17:54 Crunchums wrote:
yeah, I actually think that the best statistic for map balance is the % of the time each race was sent out on the map during proleague
i'm sure you could do even better if you did something like account for the racial composition of the entire proscene / teams / consider the race requirement

On March 28 2011 18:01 Cedstick wrote:
There's no WAY Toss are losing to Terran that much on Python. Am I looking at this thing backwards?

On March 28 2011 18:05 milikan wrote:
+ Show Spoiler +

1) won't that show up in the w/l stats, as one matchup will be skewed for that map? or does that mean sample size will be too small?
2) do you mean that a lot of mirrors are played thus leading to a smaller sample size?

in general i dont see a problem with using w/l stats as a general measure of balance; obviously its not going to be perfect, but as rough guildline i don't see why not. balance makes it easier for one race to win vs. another, so spread over a large sample size, shouldn't that be reflected in the w/l stats?

"The reason I made these graphs are to show how certain maps really are quite good like Fighting Spirit since it actually seems to converge to close to 50% for some reason, so it's a good overall signal." this point seems to agree with what im trying to argue; that a 50% w/l ratio for all MUs infers a balanced/good map.

unless "balanced" doesnt equate to "good"?

On March 28 2011 18:12 Milkis wrote:

1) Basically suppose you have a map pool that is PvT imba. Protoss eliminates all the Terran so all we have left is a lot of PvZs. Basically it doesnt give you a good overall picture and it requires quite a number of samples, but maps change every individual league so you dont really have anything conclusive.

On March 28 2011 20:32 DTK-m2 wrote:
Anyone want to explain to me how TvP is so Protoss-skewed? I fucking hate playing that map as Terran. Protoss has such an easy time taking a third, and I have a hard-as-balls time taking a third.

On March 28 2011 21:02 Severedevil wrote:

Which map, HBR?

HBR has a hard third for the Terran, but an easy fourth and fifth. That's pretty threatening for Protoss, on a map with only 6 bases that Protoss can ever take, one of which is indefensible after Terran takes his mineral-only fourth.

As a consequence, Protoss has to play aggressively enough that either Protoss gets to mine out both corners, or Terran can never take center. Fortunately for Toss, they get an early advantage (long rush distance + hard third for Terran) so they can afford to play it aggressive.

On March 28 2011 21:57 DTK-m2 wrote:


HAHA WHOOPS. Forgot to say which map. I was talking about Python. Protoss takes one of the other mains by making a wall with 2 pylons and a forge, and Terran can't take one of those mains because it's so out of the way.

On March 28 2011 17:57 Milkis wrote:


Yeah but that's also kind of hard cause some teams have "more zergs" than others etc. Winner's League also throws a lot of stuff off too since it operates kind of differently. It's something to look at in the future though, although I was planning on taking the noob easy way out and just fixed effects it and just look at coefficients in the long run to determine map balance.

On March 28 2011 22:10 ReketSomething wrote:
Outsider my favorite map :O

fairly balanced towards the end except TvZ...I play Z so mehz.

outsider <3

On March 28 2011 18:01 Cedstick wrote:
There's no WAY Toss are losing to Terran that much on Python. Am I looking at this thing backwards?

On March 29 2011 05:09 xxpack09 wrote:


Vulture play is REALLY strong on Python

Tornado terran is amazingly powerful, the open field leaves so much room for vultures to roam about and the long distances between bases keep the dragoons running

On March 29 2011 05:37 buhhy wrote:


Man, that python graph is super surprising... I've always felt tvp on python was ridiculously annoying for terran: easy to flank, hard third, while tvp on fs was much more bearable.

On March 29 2011 01:23 Crunchums wrote:

Winner's league games should be thrown out when considering usage rates for obvious reasons. I would also throw out all ace match games as well; while individual games might support a certain argument (eg KT not sending Flash for an ace match on a map support that being bad for terran) I don't think KT sending Flash or Oz sending Jaedong on a certain map says much about the map's balance.

You can also account racial differences inherent in the scene by not just comparing the data to a naive baseline of 50%. The data will actually be more useful not that matches are Bo7 too, which is nice.

There are still flaws (eg if a team gets 4-0'd we lose 2 games) but if you are doing a statistics based analysis of map balance I think it is the best approach.

On March 29 2011 06:27 Necosarius wrote:
I think it's pretty cool how the graphs evens out on almost all maps (ofc there are some exceptions. Polaris rhapsody lol.)

On March 29 2011 06:31 dRaW wrote:
I've always hated PvT on Python, I found it quite hard because the play-style is much different than newer macro maps (far third and very open, circuit breakers is a little more similar play-wise). I find it funny how the stats show it that terran favoured when so many terrans complain that it's protoss imba

On March 28 2011 17:33 Milkis wrote:



Also Polaris Rhapsody


screw that map

On March 29 2011 11:33 Carefree wrote:


Why don't we just do it by game then?

On March 29 2011 05:41 xarthaz wrote:
If you reject the premise of player skill being equal for all three races(which is the reasoning for not including WL or ace matches or whatever) then that also implies that no statistics at all can reflect the map balance

On March 28 2011 17:26 Milkis wrote:
Fighting Spirit

On March 30 2011 02:45 revy wrote:
Milkis, could you possibly show us the first derivative of these graphs (You'll likely need to smooth the plots or else they'll look like garbage) or a sliding window average?

That would help to show trends in the matchups at a given time. For instance destination was initially pretty imba for TvZ but sometime between december 2008 and march 2009 it started turning around and zergs began winning more often which eventually brought the MU down to ~55%.

If you have questions about how you might go about it you can PM me, I do these sorts of things for a living (though am out sick today!).

hmmm doesn't that mean sort of pull the statistics to extremes? if one map is really imba, then the stats will show it with a huge disparity in win rate, since there's a huge disparity + small sample size. however if a map is balanced, then there will be ~50% win rate + large sample size. so there's an exponential increase in "balance" as w/l stats approach 50%...? so assuming what i just said is true, FS is SUPER DUPER balanced? ^^

What about an aggregate racial balance chart, so we can see overall map balance over time (or was that already done before)?

Winner's league games should be thrown out when considering usage rates for obvious reasons. I would also throw out all ace match games as well; while individual games might support a certain argument (eg KT not sending Flash for an ace match on a map support that being bad for terran) I don't think KT sending Flash or Oz sending Jaedong on a certain map says much about the map's balance.

You can also account racial differences inherent in the scene by not just comparing the data to a naive baseline of 50%. The data will actually be more useful not that matches are Bo7 too, which is nice.

There are still flaws (eg if a team gets 4-0'd we lose 2 games) but if you are doing a statistics based analysis of map balance I think it is the best approach.

On March 30 2011 01:34 aimaimaim wrote:
Okay, I'm confused.

PvTwinration means Protoss wining against Terran or is it vice-versa?

So, the Python's Graph indicates a huge favor for Terrans against Protoss?

On March 30 2011 08:56 Z3kk wrote:


Yeah, it's ratio of P wins versus T to total games PvT.

And yes

On March 30 2011 12:29 Scaramanga wrote:
I'd love to see a graph for peaks, its the most balanced map ever isnt it?

On March 30 2011 15:37 Crunchums wrote:
fighting spirit

can you plot the same thing across all proleague games (ignoring the map)?

On March 30 2011 16:51 Crunchums wrote:
is it possible to mark the points on that last one where the map pool changed?

On March 30 2011 20:33 ghrur wrote:
What program are you using to make these?

On March 30 2011 23:09 Waxangel wrote:
u still need to fix them so X axis is # of games, not time