Notice the pattern? P > T > Z. We can do more, however. Using Grandmaster as an example, if we assume each race ought to be 57% then we can estimate mirror win-rates and subtract them. If ZvZ is 57% in GM, that would mean (vs P + vs T + 0.57) / 3 = 0.552.
Simplify:
vs P + vs T + 0.57 = 1.656
vs P + vs T = 1.656 - 0.57
vs P + vs T = 1.086
Now, if we assume that vs P is the same as vs T, we get:
1.086 / 2 = 0.543 = 54.3% win-rate for Zerg in non mirror matchups (matchups affected by balance).
We can do the same for Protoss:
(0.57 + vs T + vs Z) / 3 = 0.57.
(0.57 + vs T + vs Z) / 3 = 0.57
0.57 + vs T + vs Z = 1.71
vs T + vs Z = 1.14
Now, if we assume Protoss is equally broken in PvT as PvZ, we get: 57%. In other words, Zergs have a ~54.3 win-rate in GM in ZvP while Protoss have a whopping 57%.
Considering the sheer number of games (22711), there is a 99.99999999999999999999999999999999999999999999% chance Protoss is overpowered.
Protoss also play 2.3x as many games as Zergs do, which is clear indication that the health of the game is deterring Zerg players from even playing the game at all.
Blizzard, you are literally killing the game by enabling the Protoss imbalance to choke Zerg players out of GM/masters league. Protoss currently is 44.5% of GM and Zerg 23.5% in the region that these stats are taken from.
You are going to have a mass exodus of your best players leaving the game if this isn’t rectified soon. The trend has also been going on for over 2 years at this point. Please fix it. Protoss have been given enough free wins.
A 2% Discrepancy in GM and Top Masters is not statistically relevant.
Imbalance would be if it was like an entire standard deviation difference, such as 15% discrepancy, then that would be statistically relevant.
Zerg is a counter-punch race and also takes more macro mechanics and perfect decision making to play. It is not “noob friendly” at any skill level above Gold. Every Drone you train is a unit you can’t train. Every unit you train is a Drone you can’t train. Droning at exactly the right moment is highly important as Zerg.
It’s the opposite, the race starts being really bad for new players, starts getting good at plat and at diamond the average Z has more mmr than the average P or T, it balances until 4,3k more or less, so Z has around 1500mmr range in which they excel because their macromechanics are easier and better than T and P ones at that level.It’s not a lot harder or easier than other races, is just that Z does not require to keep the door closed as much as P and T, the scout is too good, and the economy ramps up very well against players who are not doing enough damage, something that happen in low levels all the time.
That’s obviously false given that individual matchup win-rates show a bigger disparity than 2%. The average win-rates show 2%, but they include mirror matchups that aren’t relevant to balance.
Your claim is also clearly false given the size of the sample. Sometimes I forget that most people have no clue how statistics work. Basically when a trend happens over more and more data points, the more and more likely it is an actual trend and not just a fluke. That depends on the size of the trend, too. A 2% difference in win-rate over 23k games works out to close to a 99.9999999% chance that Protoss is overpowered.
As the barcode pointed out, the matchmaker has demoted tons of Zerg players (about 1/3rd of Zerg’s GM representation) which means it’s the top 2/3rds of GM Zergs creating this 54% win-rate while it’s a bunch of Masters 1 Protoss contributing to the 57% Protoss win-rate. In all actuality, Zerg’s win-rate in GM is much lower than what this is reporting.
By the way, we can estimate how much this is, since we know the win-rate of Zergs in Masters 1 is 51.9%. If 1/3rd of Zerg’s GM representation was demoted, then the actual winrate is (2/3 * 54.3 + 1/3 * 51.9 = 36.2 + 17.3 = 53.5%.
Likewise we can estimate Protoss’ true win-rate based on all the Masters 1 Protoss being promoted into gm using the equation in the original post:
PvP = 2/3 * 57 + 1/3 * 52.7 = 55.6.
(x + x + 55.6) / 3 = 57
x + x + 55.6 = 171
x + x = 115.4
x = 57.7
In simpler terms, by estimating the win-rates of the GM zergs who were demoted to Masters 1, and Masters 1 Protoss who were promoted to GM, we can calculate the true GM win-rates as:
That means that Protoss does better if they have an opportunity to cheese.
Its pretty obvious Protoss doesn’t hold out in real competitive tournaments where you have bo3, bo5 or bo7.
Sure Protoss can maybe grab a few wins. But in the end they are unreliable.
Counter this with Zerg, which at the high end is very reliable and becomes better with skill. And you immediately see Zerg overperforming in tournaments.
The Protoss race doesn’t work at the moment and blizzard has refused to repair it.
Gateway needs to have proper scaling, so that it cant be used in a timing, but doesn’t require tier 3 to keep up. Then tier 3 can be nerfed.
But whenever someone says, Protoss tier 3 is OP they don’t realize its because Protoss needs that edge over a zerg who can easily grab bases or a terran who has better mobility.
I am afraid they may be too limited to actually reference stuff and give source.
I have not seen opening post, but I suspect it is connected to my thread from a few days ago.
Interesting… I think I know why it wasnt linked. If you go to the site at this time (as opposed to that topic from 3 days ago) you get the following data:
h ttps://burnysc2.github.io/MMR-Ranges/
Americas:
--------------------- P------T-----Z-----R
GrandMaster 59.3 59.5 60.1 63.4
Even worse for Korea:
GrandMaster 56.2 55.8 57.6 53.0
Admittedly Europe is not Zerg favored and therefore the only valid data:
GrandMaster 56.7 56.8 55.3 -
Look at my surprised face that the original post did not specify the source, the date, nor the server for which the data was coming from.
Like I mentioned, you completely ignore the fact that the ladder is a blind matchup and a best of 1 series.
Hence, you can use cheese more easily while if you tried that in a tournament, people would quickly catch on and after one game, your trick has run dry.
That’s the issue. You fail to give any arguments on it, but just repeat the same talking point.