Advertisement
Latest DotA VODs
Advertisement
Search
DotA drafting statistics
Posted by Chris "Zeyall" H. 2 years ago
Every match the same: What heroes should be banned, which ones should be picked? FBB.Fatal analysed 109 competitive DotA replays from our database and published the figures. Ogre Mage a guarantor for victory? Read the comments to understand the figures in the tableThe following three tables list the different heroes according to their main attribute and sort them with the amount of games won. A higher expressiveness have of course the figures of the heroes which have been at least played 10 games, no matter if won or lost.
Strength: Banned and overrated?
It starts with a lot of superlatives: the top three banned heroes are all strength heroes, namely Tidehunter, Beastmaster and Slithereen Guard. Slithereen Guard and Tidehunter are very special here: they have been on the drafting board every game, but mostly in the banned hero list. In figures: Slithereen Guard has been banned in 84 % of all the games covered by these statistics. Besides that, Tidehunter has one of the highest win percentage of all heroes, with 71% won game overall. On the other side, Earthshaker, Rogue Knight and Sand King, all of which count as gank heroes, are below 50 % of won games.
| Strength | |||||
| Hero | Banned | Picked | Wins | Losses | Win rate |
![]() | 20 | 57 | 28 | 29 | 49% |
![]() | 35 | 50 | 25 | 25 | 50% |
![]() | 13 | 46 | 21 | 25 | 46% |
![]() | 5 | 40 | 16 | 24 | 40% |
![]() | 85 | 24 | 17 | 7 | 71% |
![]() | 78 | 22 | 15 | 7 | 68% |
![]() | 91 | 18 | 8 | 10 | 44% |
![]() | 0 | 9 | 2 | 7 | 22% |
![]() | 7 | 9 | 6 | 3 | 67% |
![]() | 2 | 8 | 4 | 4 | 50% |
![]() | 0 | 7 | 2 | 5 | 29% |
![]() | 6 | 7 | 5 | 2 | 71% |
![]() | 0 | 6 | 3 | 3 | 50% |
![]() | 2 | 5 | 4 | 1 | 80% |
![]() | 0 | 5 | 2 | 3 | 40% |
![]() | 0 | 4 | 2 | 2 | 50% |
![]() | 0 | 4 | 3 | 1 | 75% |
![]() | 1 | 4 | 2 | 2 | 50% |
![]() | 0 | 3 | 1 | 2 | 33% |
![]() | 0 | 3 | 2 | 1 | 67% |
![]() | 0 | 2 | 2 | 0 | 100% |
![]() | 1 | 2 | 1 | 1 | 50% |
![]() | 0 | 1 | 0 | 1 | 0% |
![]() | 0 | 1 | 1 | 0 | 100% |
![]() | 0 | 1 | 0 | 1 | 0% |
Agility: Huge differences!
The most notable figure: Vengeful Spirit is the agility Slithereen Guard, being on the drafting board every game. With one difference: She gets picked a lot more frequently than Tidehunter or Slithereen Guard with 58:51 ban versus pick statistics. Broodmother seems to be one of the most hated heroes to play against: being picked only eight times (with 6:2 win statistics), and her ban stats are nearing insanity at 50 bans. Priestess of the Moon is the complete opposite here: she participated in 79 drafts, being picked in 63 of them. That's a record.
| Agility | |||||
| Hero | Banned | Picked | Wins | Losses | Winrate |
![]() | 16 | 63 | 35 | 28 | 56% |
![]() | 58 | 51 | 31 | 20 | 61% |
![]() | 8 | 36 | 12 | 24 | 33% |
![]() | 13 | 17 | 9 | 8 | 53% |
![]() | 39 | 16 | 7 | 9 | 44% |
![]() | 2 | 13 | 7 | 6 | 54% |
![]() | 5 | 11 | 6 | 5 | 55% |
![]() | 7 | 8 | 5 | 3 | 63% |
![]() | 50 | 8 | 6 | 2 | 75% |
![]() | 23 | 7 | 5 | 2 | 71% |
![]() | 0 | 6 | 3 | 3 | 50% |
![]() | 18 | 6 | 3 | 3 | 50% |
![]() | 0 | 5 | 4 | 1 | 80% |
![]() | 0 | 5 | 3 | 2 | 60% |
![]() | 1 | 5 | 2 | 3 | 40% |
![]() | 1 | 4 | 3 | 1 | 75% |
![]() | 0 | 4 | 3 | 1 | 75% |
![]() | 0 | 3 | 1 | 2 | 33% |
![]() | 1 | 3 | 3 | 0 | 100% |
![]() | 1 | 3 | 1 | 2 | 33% |
![]() | 8 | 2 | 1 | 1 | 50% |
![]() | 1 | 2 | 1 | 1 | 50% |
![]() | 0 | 1 | 0 | 1 | 0% |
![]() | 2 | 1 | 1 | 0 | 100% |
![]() | 0 | 1 | 0 | 1 | 0% |
![]() | 0 | 1 | 1 | 0 | 100% |
![]() | 1 | 0 | 0 | 0 | - |
Intelligence: Percentage record, but mostly below 50%!
Most intelligence heroes seems overrated: only three of the top 10 picked intelligence heroes have a winning percentage higher than 50. Storm Spirit is the hero with the worst figures overall (with more than 10 matches played): only eight out of 30 games won. But here's quite a surprise: Witch Doctor has the highest win percentage overall, and statistically speaking, three out of four games featuring Witch Doctor have been won.
| Intelligence | |||||
| Hero | Banned | Picked | Wins | Losses | Win rate |
![]() | 44 | 54 | 21 | 33 | 39% |
![]() | 35 | 49 | 23 | 26 | 47% |
![]() | 23 | 41 | 15 | 26 | 37% |
![]() | 21 | 30 | 8 | 22 | 27% |
![]() | 8 | 29 | 14 | 15 | 48% |
![]() | 10 | 28 | 17 | 11 | 61% |
![]() | 39 | 25 | 16 | 9 | 64% |
![]() | 21 | 20 | 8 | 12 | 40% |
![]() | 2 | 17 | 9 | 8 | 53% |
![]() | 8 | 17 | 7 | 10 | 41% |
![]() | 34 | 15 | 5 | 10 | 33% |
![]() | 0 | 14 | 10 | 4 | 71% |
![]() | 2 | 14 | 6 | 8 | 43% |
![]() | 0 | 12 | 7 | 5 | 58% |
![]() | 8 | 12 | 9 | 3 | 75% |
![]() | 4 | 11 | 7 | 4 | 64% |
![]() | 0 | 10 | 6 | 4 | 60% |
![]() | 1 | 9 | 6 | 3 | 67% |
![]() | 0 | 9 | 9 | 0 | 100% |
![]() | 0 | 9 | 5 | 4 | 56% |
![]() | 0 | 8 | 4 | 4 | 50% |
![]() | 0 | 7 | 2 | 5 | 29% |
![]() | 5 | 7 | 4 | 3 | 57% |
![]() | 1 | 5 | 4 | 1 | 80% |
![]() | 0 | 4 | 0 | 4 | 0% |
![]() | 0 | 4 | 2 | 2 | 50% |
![]() | 3 | 3 | 2 | 1 | 67% |
![]() | 0 | 2 | 2 | 0 | 100% |
![]() | 0 | 2 | 1 | 1 | 50% |
![]() | 0 | 2 | 2 | 0 | 100% |
![]() | 2 | 1 | 0 | 1 | 0% |
Top Ten Banned Heroes: Not intelligent
There is no intelligence hero in the five of the most banned hero, but three strength and two agility ones. As already mentioned, Slithereen Guard appears in every watch, but mostly as a banned hero. Is he just so hard to play against or is it general hate towards him? Figures show that he was only able to win 44% of the games where he actually got played.
Placement two to five have a reason to be in the top five of the most banned heroes: all of them have a winning percentage way above 50%. Broodmother and Tidehunter with 75% respectively 71%.
With four strength, 3 agility and 3 intelligence heroes, the top ten of the list of most banned heroes seems very even.
| Top Ten Banned | |||||
| Hero | Banned | Picked | Wins | Losses | Win rate |
![]() | 91 | 18 | 8 | 10 | 44% |
![]() | 85 | 24 | 17 | 7 | 71% |
![]() | 78 | 22 | 15 | 7 | 68% |
![]() | 58 | 51 | 31 | 20 | 61% |
![]() | 50 | 8 | 6 | 2 | 75% |
![]() | 44 | 54 | 21 | 33 | 39% |
![]() | 39 | 16 | 7 | 9 | 44% |
![]() | 39 | 25 | 16 | 9 | 64% |
![]() | 35 | 50 | 25 | 25 | 50% |
![]() | 35 | 49 | 23 | 26 | 47% |
Top Ten Picked Heroes: Not at all a winning guarantee
One figure seems very irritating: only two out of the top then most picked heroes have a winning percentage above 50. Only the Priestess of the Moon and Vengeful Spirit have 56%, respectively 61%.
| Top Ten Picked | |||||
| Hero | Banned | Picked | Wins | Losses | Win rate |
![]() | 16 | 63 | 35 | 28 | 56% |
![]() | 20 | 57 | 28 | 29 | 49% |
![]() | 44 | 54 | 21 | 33 | 39% |
![]() | 58 | 51 | 31 | 20 | 61% |
![]() | 35 | 50 | 25 | 25 | 50% |
![]() | 35 | 49 | 23 | 26 | 47% |
![]() | 13 | 46 | 21 | 25 | 46% |
![]() | 23 | 41 | 15 | 26 | 37% |
![]() | 5 | 40 | 16 | 24 | 40% |
![]() | 8 | 36 | 12 | 24 | 33% |
Conclusion: Math+DotA= not true!
After reading through all of these figures and statistics, one may wonder why he didn't win his next game; although he has picked the ten heroes with the highest win percentage.
Well, DotA is not math. Figures are just figures, and one should not take them into consideration while picking. To relativise the low percentage of for example Storm Spirit: maybe the player of that hero just sucked. Or his team. There are a trillion reasons for a loss, it's not just that the hero is underpowered.
DotA is so much more than just simple math and figures. It's laning, teamplay, timing and a couple of other things. But why not trying the Keeper of the Light more often? Nine games and zero loss says more than thousand lines of text and tables.
| Links |
| PlayDotA.com - Forum |
| Comments \ Paging |
1, 2
|
| What replays exactly are these? TA picked only ONCE? Storm 8/30? Hard to believe. But nice job. Must have been a bitch and time consuming to do this. |
| #42 Actually I was done before your thread appeared. Contact me for future things :) #48 it doesnt matter, KotL always wins :F |
| I liked when we had this same kind of statistics for MYM Pride at mymym.com. Btw, it's important to notice that Phantom Lancer is on a 71% winning statistic. That summed up with KotL's 100% give us some ideas of which is one of the best combinations of heroes at the moment. Still, Void's stats show us how these numbers might not represent too much... (100% wins for Void is absurd!) |
| whould be nice if someone with tha skillz could make a program for GosuGamers similar to the Dz statsthing for DL |
| some awesome statistics |
| I disagree with most of the criticism of this article. 1. The title is evidently lighthearted. It's not misleading because it's clear that it is a typical catchline, put in a popular but very simplistic form. "DotA + math = True" doesn't even make sense. 2. The input of the stats is interesting and possibly a useful set of facts, but not conclusive of anything in itself. The article states this (did you read it?). |
| THX for statistics. Seems like everyone can play better against often picked heroes and every team is better in countering them. Of course, some heroes are (and will be) under/overpowered and under/overrated but imo DotA is becoming more and more balanced :) |
| i didnt know that shaker is the 2nd most picked hero tide with 71% just shows how powerful this hero is slardar with 91bans and 18 picks is just WTF |
| Dota picks work mostly like fashion. Ditto Slardar. Now he's in, a month from now he'll be out. It's not as if he received any major buffs or anything. |
| My Conclusion -Mass AOE stuns (mainly streanth) and disable heroes dominate this version.Special mention to ES who has exceptional early and late game and tide who, as of this version, is imba.Also Slardar could see a small hp/armor nerf. -Heavy late game DPSers don't have any place in this version cause they need too much farm to do anything early and suffer from mass disable/guinsoos late game. -Agi heros with genuine pushing ability being favoured (can push lanes/towers, farm better AND have the advantage in late game -Most teams are too faggish to try new strats. That's why mirana (not exceptionally good but jack of all trades) and venge (kinda imba this version) are getting overpicked. |
| kotl 9/9 andd put pugna instead of sf 39 times picked sf only 36 ! |
| is interesting. cept the sample size is way too small =/ |
| Thx for this. Ive been wanting to do this myself for a long time tho the truth is that it might be too hard to do this accurately as the metagame constantly changes, for example, you would definitely not get those kinda win % 3 months ago. |
| really nice, i liked it gj |
| nice article! |
| actually a sample size of 30 is suffecent for most statistixal analyses. there are just to many lurking variables within the data to make it have any significance |
| nice read+numbers =P |
| Wow good job ! |
| Where are nerubian assassin and dark seer on the top 10 banned list?!?!? |
| I agree this sample size is to small. You will need many many more games to get an accurate reading off the statistics. |
| Now, IceFrog will have something to ponder over ;). |
| they should enter dota in the curiculum hahahha |
| Mirana is the most loved hero of all times ,nothing new there,replay parsing sites of old had huge amount of games that pointed this way :D. |
| Good Work, keep on updating it :D |
| I'm surprised that Nevermore was in the top 10 picked heroes... |
| Brilliant idea! Although not enough observations; and also have to account for difference in map versions. Time-series regression. Nice initiative, some plz do a census? :P |
| These statistics are incomplete as some heroes are picked to counter something, otherwise they would be useless. If more info was added here, it would become an article about picking tactics, not statistics. This is what i like DotA - its complex and quite balanced, no certain wins included. |
| Well, DotA is not math. Figures are just figures, and one should not take them into consideration while picking. To relativise the low percentage of for example Storm Spirit: maybe the player of that hero just sucked. Or his team. There are a trillion reasons for a loss, it's not just that the hero is underpowered. DotA is so much more than just simple math and figures. It's laning, teamplay, timing and a couple of other things. But why not trying the Keeper of the Light more often? Nine games and zero loss says more than thousand lines of text and tables. TOTALLY BULLSHIT - dota is all about who farms an agility hero on 50 min. |
| wow that's amazing |
| This means absolutely nothing. You wasted your time. |
| nice |
| #3 maybe u should learn about statistic lesson... that using sample 100 from 1000 games / decided 10% population is already enough, and i think this statistic from nowadays game KOTL is not guarantee, but TreantProtector is guarantee... this hero is annoying when he got RO + dagger.. |
| Guarantee is bullshit, if it would be guaranteed to win with a hero, it wouldn't be a game anymore. It would be like playing a computer who doesn't make any mistakes and his decision making would be flawless, and even then you could do the same, but guarantee = no need to play. As for the statistics, yes with 100 matches it's ok, but like I said it depends on the skill level of the teams. It would say more if you pick final matches of BIG TOURNAMENTS and compare those, which would be hard to get 100 matches. If my clan plays against EHOME I'm sure we loose 100% and that's not even statistics, but the heroes picked wouldn't change anything. So if they play the "worst heroes" possible they would win and still the win% of those heroes would rise in THIS analysis. |
| This chart gives us a good idea of the current metagame [of games posted to Gosugamers, which doesn't include a huge # of games especially Asian ones], but the fact that data from both -cm and -cd games are pooled makes it pretty hard to draw any strong conclusions about MOST of the heroes (there are some exceptions). Whether the data itself is flawed or not is irrelevant to my next point which is that the conclusions drawn by the writer are flawed as well. You can't look at the list of the top 10 most picked heroes and say that their <50% win is "very irritating" because you would expect all of these heroes to be <50% win... the fact that potm/vs are above 50% and they are not in the top10 bans is what's most surprising. Heroes that are picked most are also played against most and they lack "surprise factor" so teams are prepared to play against them, which is where counterpicking and teamplay come in as major factors. In an idealized scenario I would expect all of the top10 picked heroes to have <50% win. |
| nerf kotl imo... definitely. dark seer is extremely underpowered guys. |
| Legatoo mentionned another argument I guess... the fact that a hero is picked often gives training to the enemy team so they will play better against the next game. Furthermore this doesn't mean that they stop picking that hero because it's still a good hero which is maybe always nice to have on the team. |
| Firstly good job for the article (how long did it take you to gather all that statistics?:)) ) However, I have to agree with #3 that the sample size is indeed not that big. PS: lol I thought Levi always loses :) |
| must learn from puppey... |
| I dont know what this is good for... |
| Hmm doesnt say anything cause the skill-lvl is in most of the games very different between the teams. Even with crap heroes especially when u look at europe/america their top teams like mym, blight (nirvana), ddt defeat other teams cause they play totally different lvl. Really stable close teams in europe what promise good matches are rare. But nice is this statistic anyway! |
| #43s right . nice analyze.tnx for share |
| tide is good... nothing else can really be deciphered from these stats. should analyze more games in a global scope |
| #23 So true man, so true. |
Advertisement
DK
vs.
Tongfu
16h+
DT.CLUB
vs.
TyLoo
40h+
iG
vs.
NA
64h+
LGD
vs.
WE
112h+
HOW
vs.
Gx.Brutal
138h+
iDeal
vs.
Neolution
138h+
ECS.Fight
vs.
Trust
138h+
Neolution
vs.
HOW
162h+
iDeal
vs.
ECS.Fight
162h+
Latest from the forums
Introduce.
(15)
Poll
Which Chinese team would you most likely go BANANAS over if you saw them in a video interview by GosuGamers? (744 votes)
DK | (53%) |
||
iG | (21%) |
||
LGD | (9%) |
||
Other | (7%) |
||
Panda | (2%) |
||
WE | (1%) |
||
Tongfu | (1%) |
||
TyLoo | (1%) |
||
For Love | (0%) |
||
Most recent blogs
May 27
blitzking: Starting with Gosu Gamers
(02)
May 24
ZEEERYX: 3422 Championship#DOTA2
(00)
May 22
dota_woohoo: Need 3 dota2 players
(04)
May 21
vaneov: Deaf Dota 2 Open I
(02)
May 21
yOx: The pANDA mist
(01)
May 19
Ferrar1: Mvp or Squirtle GSL S2
(01)
May 19
Ferrar1: International 2012 Update
(01)
May 19
Ferrar1: New CLG and Zenith Dota2
(02)
May 19
Pianoparanoia: Mineski Week (for me)
(03)





























































































1
You are not a logged in, or you are not registered. Create an account or login to be able to post a comment.














