|
On April 26 2011 02:04 JJH777 wrote: Nestea said this: Terran you can be good after only playing a little, protoss you have to work hard but once you do you are unbeatable race, zerg sad.
I think that fits those statistics decently. Is it necessary you post something almost identical in most of your posts?
Actually, a lot of people are making nearly the same exact post, and its getting old fast.
I could just as easily say "MVP recently said in an interview that terran is underpowered"
Does that contribute anything?
|
On April 26 2011 02:04 JJH777 wrote: Nestea said this: Terran you can be good after only playing a little, protoss you have to work hard but once you do you are unbeatable race, zerg sad.
I think that fits those statistics decently.
nestea said terran can be very good after practicing hard
then he said protoss will be unstoppable after practicing hard
+ Show Spoiler + [.IMG] [./IMG] around the link to show pictures without the .s of course
|
Does it actually matter how many people are playing "race X" on the ladder?
It doesnt have any significance for balancing the game. It is only important for the "fun factor" when viewing tournaments and such, because tournaments with only 2 out of 3 races are less exciting. There isnt a significant skew in balance among the top tournament players however and thus the race distribution seems to be ok.
If players only have mirror matchups on the ladder it is their own fault, because they are playing the most powerful / easiest to play / prettiest race. Well its actually not entirely true, because Blizzard shares part of the blame due to the horrible tiny ladder maps and their close ground spawns, which disfavour Zerg and which make rushes against Zerg too easy for Terran and Protoss. Thus they are responsible for making these two races more more powerful on the ladder. Luckily most tournaments are using the sensible and much bigger GSL maps (and some iCCup ones as well).
|
On April 26 2011 02:17 TheBB wrote:Show nested quote +On April 26 2011 02:00 buldermar wrote: Firstly, I want to point out that I'm taking a mathematical approach to my question.
(...)
Based on the above statistics, I feel safe to make the conclusion that terran is dominating the top of GM League in every way I could think of. If you really want to take a mathematical approach, you should work out some p-values before coming with a conclusion, or nobody will take you seriously.
It is somewhat tricky to do statistical significance testing (at least with this method) due to my samples interrelation. For instance, while doing it on a samplesize of a dice, each throw does not influence the next throw, but when looked at the top rankings, two issues arise;
a) Each spot has a certain level of significance for the overall picture, the number one spot has the highest and the 100th spot the lowest (in the case of points top 100)
b) When dividing the spots into sections of 10, which is nesseccary unless I want to do a huge amount of work in excel OR give an overall picture of top100 instead, what shows in one section influence the likelyhood of each race being represented in the next.
There are ways to go around this, but it's time consuming. It would be a lot easier, but also a lot less interesting, to look at, for instance, Just "top 100" of GM ín relation to all of the players in GM.
Finally, who are you to decide the level of seriousness for the entire community?
|
On April 26 2011 02:17 Shaetan wrote:Show nested quote +On April 26 2011 02:06 buldermar wrote:On April 26 2011 02:02 Shaetan wrote: You can't claim to do statistical analysis then just throw out numbers and claim that X is true, you have to do the analysis. Can you examplify? I thought I did a decent job at showig a trend based on every aspect of the top of the ladder I could think of. I even specifically pointed out that I'd gladly look into more if anyone got ideas. You need to show that the difference you see is statistically significant and not just due to chance. You said yourself when looking at a sample size of ~1600 win rates were much more even so it's possible decreasing your sample size artificially created the disparity. Also not sure why you wouldn't look at the full 200 or at least all GM league top 100.
Top 200 includes too high an amount of players to represent the top of the ladder. When looking at only top10/top20 etc, the outcome can be explained solely with chance, but when looking at multiple areas you'll have to multiple the respective odds of it being due to variance for each area, which changes the picture dramatically. The issue is that many of the areas interrelate, making traditional statistical significance meassurements difficult.
My only excuse for not looking at entire top 100's the way I initially did is it being time consuming. I decided to just look at top 30 for each of the other groups as the top spots has the highest level of significance for what I'm looking into.
|
On April 26 2011 02:18 ForTheDr3am wrote:Show nested quote +On April 26 2011 02:12 Cryllic wrote: Well yea these would apply mostly to ladder games. MVP said that terran was the weakest race in GSL maps because of the large rush distances and lack of early game aggression which is where the terran race excels at.
This is true. Especially if some people try to imitate builds that they are going to play on the bigger GSL maps on the smaller ladder maps, it can be very easy to randomly lose games against Terran.
But not for terran to randomly lose games when doing so, because their openings are less dependant on the size of the map, or?
|
On April 26 2011 02:19 Shaetan wrote:Show nested quote +On April 26 2011 02:10 buldermar wrote:On April 26 2011 02:06 tehemperorer wrote: Well, from a Protoss perspective, if you are in GM and you get a lot of PvP matches in one day, it seems you could really either skew your score or drop in points dramatically because the matchup is very volatile (coinflip BO loss people tend to say).
To be a top player you have to be dominant in every matchup, not just 2/3, and since PvP isn't figured out yet, top players might still go 50/50 in that matchup, whereas a Terran player can be really strong in their mirror and can win them most of the time? It seems that the Terran mirror is the best mirror because of the differing ways it can be played and still won. In other words you're stating that for any given TvX matchup, the terran player will be much more likely to win asuming slightly superior skilllevel than what would be the case for PvX and ZvX matchups? Can you elaborate on this? He's saying TvT more dynamic than other mirror matchups. Although I'd say ZvZ is similar. PvP is pigeonholed into 4gate. But I'd say mirrors moreso than other matchups let the best player win since "imbalance" is nonsensical.
Yes, but that's not because imbalance is nonsensical, that's because more boils down to micro and less is decisionmaking/openings (since openings are more commonly exactly the same OR equally good, i.e. 4gate vs 4gate). It is in other words very hard to lose a mirrormatchup of PvP if you simply 4gate and have 200 APM compared to your opponents 100, and the same applies to TvT in other ways. The most luckbased mirror matchup would be ZvZ since multiple common openings are more clear counters to other opponents (12p>9p, 9p>15h etc)
A simple way around this issue would be to always have BOX's and let the outcome of this best of decide who lost points and who won, but that means players will have to ahead of time estimate if they have the time to play X amount of matches, so it's not likely to be implemented despite clearly improving on the correlation between skilllevel and likelyhood of winning.
|
On April 26 2011 02:20 Apolo wrote:Show nested quote +On April 26 2011 02:06 tehemperorer wrote: Well, from a Protoss perspective, if you are in GM and you get a lot of PvP matches in one day, it seems you could really either skew your score or drop in points dramatically because the matchup is very volatile (coinflip BO loss people tend to say).
To be a top player you have to be dominant in every matchup, not just 2/3, and since PvP isn't figured out yet, top players might still go 50/50 in that matchup, whereas a Terran player can be really strong in their mirror and can win them most of the time? It seems that the Terran mirror is the best mirror because of the differing ways it can be played and still won. That's how MC has above 90% win rate in PvP, right?. For sure it isn't 50/50, the best player will have the upper hand. A good player must be good in all matchups, and you can't say that PvP win is based on luck. Even then it's only 1/3 of all the matchups, even if it puts the win ratio closer to 50% it shouldn't affect as much as we see. Protoss from being the most played race, yet having the least players in top.
I agree that it should not affect itas much as we see, but I want to point out that the winrate you are referring to is from tournaments afaik, and they use a different map pool with the avarage map being significantly larger, making it much more skillbased since blind all-ins are less effective on avarage and games are longer --> more decisions to be made each game.
|
On April 26 2011 02:24 strongandbig wrote: First, the sample size of [top10] or [top30] is not big enough that you would expect it to be equal in the absence of imbalance. Over a thousand players, you can safely assume that the distributions of skill and race choice are independent. However, with ten players it's not unreasonable that there are just more Terran players of high skill.
Additionally, this assumes that winning enough on ladder to get into the top ten is dependent only on skill and game balance, whereas I think that luck is also involved; it's easy to imagine a player whose skill is #3 losing a lot of points in a BO loss to a player whose skill is #150.
Honestly OP, I just don't think you have enough data to say anything here.
The individual samples are not of large enough size, but put together it's different. I would happily look at thousands of players but the disparity in skilllevel would be too significant for the outcome, that's why I've had the neccessity of dividing players into groups of ten in my observations.
Keep in mind that despite only looking at smaller groups at once, they're all from the same source of candidates which would be the entire GM league. Therefore, it's certainly unreasonable to think that there's simply more terran players of high skilllevel - at least to this extent.
Of corse luck is involved for any individual player or any one throw of a dice but when looking at many areas in different ways and they all show the same overall tendency, luck becomes an inadequate explanation.
|
On April 26 2011 02:25 dogmeatstew wrote:Show nested quote +On April 26 2011 02:00 buldermar wrote:As you already might notice, top 6 spots is being occupied by terran players. However, a larger samplesize is required to show trends: Top 10; 7 terrans, 2 zergs, 1 protoss Top 20; 12 terrans, 4 zergs, 3 protoss, 1 random Top 30; 14 terrans, 7 zergs, 8 protoss, 1 random Top 40; 16 terrans, 12 zergs, 11 protoss, 1 random Top 50; 23 terrans, 12 zergs, 14 protoss, 1 random Top 60; 27 terrans, 14 zergs, 18 protoss, 1 random Top 70; 31 terrans, 17 zergs, 21 protoss, 1 random Top 80; 34 terrans, 19 zergs, 26 protoss, 1 random Top 90; 37 terrans, 22 zergs, 30 protoss, 1 random Top 100; 41 terrans, 26 zergs, 33 protoss, 1 random http://img853.imageshack.us/i/ranksracestop100gm.png So based on this graph, over the course of the top 10 through top 100 players it actually seems fairly linear to me, the ratio of terran to protoss to zerg players at the top is consistent through these intervals which doesn't seem to suggest imbalance at all but would rather suggest that there are just more terran players. I'm not sure why you're reading "more terran at every level" as "imbalanced" when its a linear progression at a relatively consistent ratio for all races through the top 100...
I'm not, and it's only a linear progression at a consistent ratio for all races through the top 100. When you look at the entire GM league the picture is different as terran only takes up about 30% of the spots.
On April 26 2011 02:25 dogmeatstew wrote:Show nested quote +My first hypothesis was that more people were playing terran in GM league. That would be a logical explanation, as there will naturally be a clear correlation between the amount of players playing each individual race and the percentages of each individual race in the top rankings. However, is this not the case. Here is the race distribution by league graph taken from sc2ranks.com: http://img687.imageshack.us/i/racedistributionbyleagu.png I don't like that you jump from "I'm going to only going to look at GM league because its the most accurate representation" to make your first point then switch straight to "looking at all the leagues to disprove a reasonable hypothesis".
I don't "jump to looking at all the leagues", I'm still only looking at GM league.
On April 26 2011 02:25 dogmeatstew wrote:Show nested quote + There currently is 1638 players listed in GM. The distribution is: 2.6% random 38% protoss 30.3% terran 29% zerg
For those 1638 players, the avarage win% are: Random: 52% Protoss: 56.6% Terran: 57.8% Zerg: 56.3%
For those 1638 players, the avarage amount of points is: Random: 249 Protoss: 325 Terran: 353 Zerg: 335
Looking at the same distributions of only top 100 players in GM we now have; 1 random 33 protoss 41 terrans 26 zergs (You might recognize these numbers from before)
For those 100 players, the avarage win% is: Random; 74.1% Protoss; 66.6% Terran; 68.8% Zerg; 66.5%
For those 100 players, the avarage amount of points is: Random: 657 Protoss: 661 Terran: 683 Zerg: 668
As you can see, terran is once again dominant in every area.
All the win %'s are very close. The average number of points is very close, 15 points isn't a lot and as you've already stated, a large amount of top 10 is terrans which would easily skew this. In short I'm not sure how these numbers give you the impression that terran is dominating in every area, a 2% higher win ratio isn't statistically significant even across 1638 players and all in all this last set of data actually looks pretty even across the board to me.
I agree that this part is less important due to it being influenced too much by the top rankings, which we already looked into. However, looking into this as a seperated sample, the gap is quite big. We are looking at 1638 players where terran is ahead of zerg with 1.5% winratio and protoss with 1.2%. In terms of avarage points, terran is ahead with 18 of zerg and 28 of protoss. I'm up for looking into this seperated from what else I've observed. I should have written down the exact amount of games this sample is based on yesterday, but I can tell that it is 294415 right now. With a number this high, the respective gaps cannot be explained solely with variance.
Anyway, what I'm mostly interested in is the top spots, not GM altogether, and why these topspots are being overrepresented by terran players.
|
On April 26 2011 02:29 Let it Raine wrote: as a zerg player who likes to say x is op
ladder doesn't mean much.
Ok? Did you read the post before replying? Are you aware that I'm not trying to estimate the correlaton between ladder and skilllevel but solely trying to answer why the top spots are being occupied by mostly terran players? I'm trying to figure out if this is due to map, certain matchups, the point system, lack of metagame etc etc.
|
On April 26 2011 02:34 Rabiator wrote: I dont think the number of points is a good thing to analyze simply because the amount of points you get for a victory is somewhat undefined. Too much depends upon your opponents stats and the win-ratio is a much clearer stat which says more or less the same.
Generally I dont think any result of "more players of race X at the top" says anything about the state of the balance of the game. Even the top players chose their race a year ago and not many of them change race, so any of those who do change race will need some time to get to the top of the new race. Practice and experience keeps the same players at the top.
Good thing I'm not looking into the state of the balance of the game then.
|
On April 26 2011 02:34 FabledIntegral wrote: Can someone explain to me why there are 1638 players in GM? I wasn't even aware there were 8 servers total.
I apologize if this has already been answered.
Every region has more than 200 GM players atm, Russia with the highest, being 211. (as of today with the total of GM players being 1644)
|
On April 26 2011 02:36 Euronyme wrote:Show nested quote +On April 26 2011 02:04 buldermar wrote:On April 26 2011 02:02 udgnim wrote:Why does terran crush the top ladder? they are the most resilient race to random strategies / attack timings? Does this apply specifically to ladder games? In that case, how does the ladder map differ from those being used in tournaments like TSL, GSL, Dreamhack, NASL etc? On the ladder people tend to do random weird stuff that they come up with themselves alot more than what happens in tournaments. Sure there are some things in tournaments such as foxers mass marine, thorzains thorbuild, guineapigs void ray collosus etc, but in ladder there are a bunch more random attack timings, and if you're playing a resilient race that's easier to defend with, you're gonna come out on top alot more often. Also on the ladder you face new opponents every time, so a weird 15 minute proxy hatch (to make an example) can be very successful, whereas in a tournament bo5, you're simply not gonna win 3 games with the same cheesy (cheesy as in out right cheese, or just not solid) strategy. TL;DR, Ladder = cheese, and terran's good at stopping cheese. The same cheddar doesn't work in a bo5.
You're contradicting yourself.
First you point out that cheese is more frequent on ladder and terran has an easier time defendin against it, improving on their avg winrate. Then you point out that cheese is less viable in a BO3+ when you have already pointed out that it isn't viable against terran in a BO1 on ladder?
|
On April 26 2011 02:38 Supah wrote:Show nested quote +On April 26 2011 02:25 HardCorey wrote:On April 26 2011 02:20 Apolo wrote:On April 26 2011 02:06 tehemperorer wrote: Well, from a Protoss perspective, if you are in GM and you get a lot of PvP matches in one day, it seems you could really either skew your score or drop in points dramatically because the matchup is very volatile (coinflip BO loss people tend to say).
To be a top player you have to be dominant in every matchup, not just 2/3, and since PvP isn't figured out yet, top players might still go 50/50 in that matchup, whereas a Terran player can be really strong in their mirror and can win them most of the time? It seems that the Terran mirror is the best mirror because of the differing ways it can be played and still won. That's how MC has above 90% win rate in PvP, right?. For sure it isn't 50/50, the best player will have the upper hand. A good player must be good in all matchups, and you can't say that PvP win is based on luck. Even then it's only 1/3 of all the matchups, even if it puts the win ratio closer to 50% it shouldn't affect as much as we see. Protoss from being the most played race, yet having the least players in top. MC is an extraordinarily good player, and outlier one may say :D. PvP is still very volatile in say low masters vs. high masters or even above that. MC's micro is on an entirely different level than pretty much every other Protoss out there. As a Protoss player, I think Terran dominance has a lot to do with both the stability of Terran in regards to cheese and in regards to a mirror. A Terran can wall in, repair the wall in, and defend the wall easiest of the three races. All their units are ranged, SCV repair is extremely powerful and efficient, and they can remove the wall with no cost.I think being the closest to cheese proof is a huge boost given the nature of ladder and those that play in it. ZvZ and PvP are hugely oriented toward the Early to Mid (rarely) game. And even if they get past T1, the timings are thrown way off compared to standard play. TvT is helped by the weakness of the Marine against bunkers/an early Siege tank. All in all, I don't think Terrans are necessarily OP or any race UP, but as the metagame is figured out atm, they are definitely the most stable and able to play a diverse style.
You're stating that terrans are not OP on the ladder AND are having an easier time winning games in which there is a skillgap than any other race. This is contradictory when looking solely at the ladder and how it works.
|
On April 26 2011 02:48 Poopi wrote: I think it has to do with : - the mappool is not as disfavored for terrans on ladder as the GSL/tournament's one is. - in BO1 that doesn't matter much, aka ladder, I'm sure a ton of ppl try some agressive / risky or even all-in strategies, and terrans are relatively safe in tvz, not so much in tvp (like the void ray pushes we have seen MC or Alicia do, and other stuff), and they can themselves do strats they can't do in bigger maps. - not sure about this one, but I heard pvp is volatile, and zvz a bit, whereas tvt is much much less volatile. - THIS IS LADDER (what I mean : MarineKing would be REALLY really happy if being 1st on ladder for the longest period a time gimmes you a title...)
As of now - with what I've taking into consideration - the current mappool seems to play the biggest role in terms of terran being the most dominant. I know little about TvT, but I know about the mapsituation from the perspective of both protoss and terran. In particular, from zergs perspective, at least one map seems always unfair in nonmirror matchups (delta quadrant due to vunerable expo), and some seems to be unair sometimes (slags pit close pos, metalopolis close pos, shattered close pos etc).
If we add to this hypothesis that TvT migt be less volatile than other matchups, and TvX might do better against frequent cheese overall, making the more skilled player in a game with a skillgap much more likely to win than what is the case for other matchups, we might have put together a possible explanation.
I wonder if there is a way to verify or falsify this.
Thanks for being so direct in your answer.
|
Stop bumping your thread. You can easily fit all your replies in a single post.
My rage towards misuse of statistics are akin to when muslims witness Quran burnings. Stop it, numbers are sacred objects.
|
On April 26 2011 02:57 zarepath wrote: Poopi has a lot of good reasons. Also, the best ladder players are not necessarily the best players in general. Winning a large number of Bo1s denotes skill, but not the most accurate measure of skill (which would probably be tournaments with bo3s and larger).
Terran is also probably the best race for Bo1s. They can easily adapt and can easily be aggressive. Someone said "most resilient to cheese," which is a good insight, too.
You can use this data to argue that Terran is the best race for the ladder, but the ladder is not the benchmark for ultimate skill, I think. Perhaps Protoss and Zerg are better for BoX matchups for some reason?
Yes, I'm inclined to think that both zerg and protoss would perform significantly better against terran on larger maps. I don't have much insight in the metagame perspective and I'm yet to find a reason to think that would in particular favor zerg/protoss over terran.
We can consider large maps compared to smaller maps for TvZ and TvP;
In TvZ, terran has the option of relying on harras. This can mean 2rax opening, 8rax 1base AI, early reactor hellions, blue flame hellions, banshees, blue flame hellion drop, stim marine drop, marine+scv AI etc. In common for all of these is the fact that they're less viable on larger maps due to the longer travel distance. This means that zerg can safer play a macrogame in which a) the more skilled player is more likely to come out ahead because more decisions are being made in a longer game and b) they can rely less on scouting information for the beginning of the game which makes getting that overlord into the base or seing that factory with poking zerglings at the ramp less important. It simply has less impact on the outcome of the game.
In TvP, protoss simply has the option of proxying a pylon when pushing, effectively negating any influence of the mapsize when they're being aggressive. On this same token, terran does not have this option and therefore their push will suffer from having to travel a longer distance, giving protoss more time to prepare. Similar, when harrasing with a warp prism, you're not inclined to have units in it. You can simply warp them in when you need them, where you need them, whereas a terran player will essentially have to remove the units from his army that he wishes to harras with. That means the units will not be a part of his army for a longer period of time on larger maps.
There is obviously much more to look into, but I can see why some people argue that larger maps favor Z and P more than T in TvX matchups, and maps being played in tournaments are larger on avarage than maps being played on ladder.
|
Easy, with such a small sample size you could just have better players playing terran.
|
On April 26 2011 02:58 Nerski wrote:Show nested quote +On April 26 2011 02:11 buldermar wrote:On April 26 2011 02:08 lilky wrote: So basically... Races in order of most played to least played: Protoss, Terran, Zerg
Races in the Top (Insert any number here) in order of most to least: Terran, Zerg, Protoss
TLDR: Protoss is underpowered.. WHAT?!?!?!!?!?!? Correct, that's my overall conclusion based on the observations I made and I'm open for expanding on those observations if anyone thinks I'm being biased. So what you're saying is this is a balance QQ with a clever 'statistical look' to it? Someone lock this sham of a thread, you're not even using statistics correctly just listing out the fact there is more terran on top then toss or zerg. For any statistic to be truly significant the deviance needs to be high enough. Not to mention GM doesn't even represent loads of the actual top players, just the top players who ladder a lot. So there is plenty of margin for error in any statistical breakdown of the GM league.
No. As I believe I've mentioned, my intentions with the thread is to cast light on what could be the explanation to why terran players are currently being more dominant than other races on the top of the ladder. This is not a balance debate.
Nobody forces you to read the thread. It if doesn't fit your spectre of interest, simply open a new thread. I havn't violated any of the guidelines and was careful not to do so.
I don't believe any of the guidelines states that I have to use statistics in a certain way. I also believe that I am free to discuss something even if it leaves room for a margin for error in it's statistical breakdown.
|
|
|
|