GGL Wire » Post: 'One year of WC3 stats: Orc and Undead the weakest races'


One year of WC3 stats: Orc and Undead the weakest races

The most imbalanced match-up in Warcraft 3 is Human (and not Night Elf!) vs. Orc, if you are to trust statistics. If you are to trust statistics, Manuel “Grubby” Schenkhuizen should hope to meet Night Elf players only if he wants to win the World Cyber Games. GGL Wire has counted!

The question of imbalance in the most popular RTS game worldwide is as old as the game itself. In order to get to the bottom of the problem, we have decided to collect the statistics from the deciding stages of professional level Warcraft 3 tournaments in the last 12 months, starting with the 2006 WCG finals.

The findings were interesting to say the least.
The aim of the experiment was to collect enough data in order to get a good impression of the balance in modern competitive Warcraft 3 at the highest level. It was not to make the most detailed and fastidious account of the game’s history. We got the stats for 334 map wins (excluding mirror matches) from the following tournaments:

  • BattleNet Season IV finals,
  • BattleNet Season V finals,
  • Warcraft 3 World War Summer Grand Prix,
  • Blizzard WorldWide Invitational,
  • World Series of Video Games finals 2006
  • WSVG Wuhan
  • Game-X
  • International E-Sports Festival
  • Electronic Sports World Cup
  • IEST 2007
  • PGL 2007
  • WEG e-Stars

According to the overall statistics, the Human race is the strongest one in WC3 and Undead is the weakest. Human players won 52.4% of maps, followed by Night Elf with a 50.5% ratio. Orc players won 48.9% of maps and Undead came out victorious in 47.9% maps. So much for one race being superior to others.

More interesting were race vs. race statistics because they did not suggest that the game is all that balanced. The fairest match-ups are Human vs. Night Elf and Human vs. Undead. Hands down the most imbalanced game in modern WC3 is Human vs. Orc. Here is how the numbers break down:

Undead > Human - 50.9% (map count: 26-25)
Night Elf > Human – 51.8% (41-38)
Night Elf > Orc – 58% (29-21)
Undead > Night Elf – 58.1% (32-23)
Orc > Undead – 63.4% (40-23)
Human > Orc – 66.6% (24-12)

The conclusions here are that the Night Elf match-up is actually the most balanced for Orc players, with a 58% loss ratio. The other Orc match-ups are either “very easy” (vs. Undead) or “very hard” (vs. Human). Theoretically, Orc players will need more bracket luck than any other race to win a high-profile tournament.

On the other hand, the race that does not need a lucky draw with the opponents’ races is the Human race. Human’s weakest match-ups are below 50% but very close to it, with a “sure win” against Orc players. This lack of an apparent weakness may explain the consistency of Human players regularly performing well in competitions. Night Elf is similar to Human in this respect.

Both Orc and Undead will have players that win tournaments but also exit them prematurely. With exceptionally weak match-ups, they need luck to win events. The results in modern competitive Warcraft 3 seem to back this claim up. It may be difficult to believe, but Grubby is the only Orc player to have won a major international tournament.

Luckily, the above are only numbers that help describe reality. They do not mean that we will not see an Orc win the World Cyber Games in Seattle, but it does seem much less than likely.

Share this via:
  • Digg
  • Facebook
  • MySpace
  • Netvibes
  • StumbleUpon
  • Fark

12 comments to “One year of WC3 stats: Orc and Undead the weakest races”

  1. I really wouldn’t call this conclusive in any sense of the word.

    To properly determine imbalance in WarCraft 3 you would need to do a detailed analysis of the races, the units, the match ups, the lot really.

    This seems like it should be a back up point to a major analysis of the game, not a stand alone piece of evidence.

    This is more liable to incite and spur on flamers as apposed to encourage debate or lead to a patch.

    You never said it would though and I’m just making a point, not stirring the pot.

  2. Well it might make sense if there were 10 times as many games indexed or the same number of players from each race but this is’s very concludent in my oppinion, especially since at almost every big tournament in top16 you have about 8 elves from which 2-4 lose basically because they face way better players while 1 or 2 always reach the final.

    Still, somehow i agree with the conclusion, it’s always been known that orcs need bracket-luck to do something in elimination tournaments and knowing that WCG 2007 has maybe no undead with real chances of getting to top8, i fail to see how orcs will benefit from their strong matchup while both hu and elf improve their virtual chances of winning; moreover i don’t see any orc being able to beat 3 players that represent hu/elf in a row (assuming he got an undead in top16), just way against the odds.
    But, of course, surprises happen and warcraft is, like any other sport, so much about shape and preparation that we might see one orc player lifting the trophy and maybe an undead in top8 which would be a surprise of equal proportions

  3. make *more* sense, is*n’t*

  4. Oh no, this is by no means conclusive. In fact, I don’t think anything will be. The numbers can be gathered like that for fun more than anything else.

    But it’s GOOD fun.

  5. first off, if the numbers are only for fun there is no need for a article because that invalidates every conclusion from the start. I think the only solid evidence on this matter is to get all the results from battle.net over a longer period of time. Something wich only blizzard has access to.

    Secondly as Teg-Kramer said, you have to take into account the pool of players creating the results. I’m pretty sure that sky and tod are responsible for >50% of all the human wins. So basically you are saying that e.g. human has a 51.8% probability vs elf based on having 2 exceptionally good players. Whereas elf has 48.2% vs human for a far larger group of players. In other words if you take sky & tod out the percentage for human will drop, whereas if you take moon out the percentage for elf will pretty much stay the same.

    Lastly, “. If you are to trust statistics, Manuel “Grubby” Schenkhuizen should hope to meet Night Elf players only if he wants to win the World Cyber Games.”

    Looking at your stats, clearly he would want to meet undead players only, or?

  6. I wanted to check how it pans out at the highest level in competitive WC3 - this automatically means that I would look mainly at players like Moon, ToD, Sky and Grubby.

    Looking at stats, Grubby should want to meet Undead players - but there are NO world class UD players at the WCG. If he wants to get far, he needs to beat world class NE and HU players.

  7. this article is so true! -_- nice article btw!

  8. Carmac, then he should hope to meet first all ud’s, then all orcs. Only thing worse then Ne would be HU for him.

    So the statement about grubby wanting to meet NE players should have 2 exceptions, UD and other orcs.

    So a better funny conclusion would be that moon and sky want to meet Grubby based on statistics.

  9. very nice again carmac, but i think you should had include the top players statistics vs themselves, then it would be a VERY VERY nice article which were never been there ;)

  10. by the way, whats up with the Video Collection of Wc3//Grubby iviews? :P

  11. Get off Grubbys dick already

  12. Good article. I love WC3 and I’m highly trained with statistics. My data collection for WC3 stats has been on computer simulations, not live players. I would love to get my hands on some of the data you worked with. If you could e-mail it to me, I could do some more advanced analyses and answer some of the questions that the posters wrote.

    For the Horde!

Leave a comment