take this with a grain of salt since i haven't looked into these things in awhile...
2. fewer than 20-25% of games are ranked (edit this is outdated, look below)
3. Not sure other than the majority are AP
okay i got curious and looked at graphs.
2. Non-ranked is ~61% (it used to be higher). Ranked is ~40% (last 4 weeks)
3. AP is 87% for the last 4 weeks. AR is 11%, other modes are the rest
4. I don't have the abandonment numbers, but i do have significance as we calculate it; around 5% of matches. Actual abandons should be higher than this.
A couple years ago, if I remmeber correctly the distribution of games played was about (75% Normal) (15% High) (10% Very High) games. Even though there are few players in Very High, they are typically people who play several games a day, while Normal Skill players might only play once a week or less.
@ relentless i think this is a difficult thing to grok since the skill brackets are.. as far as i know calculated based on the skill of the current active games. So you could change skill bracket depending on the current match volume and matchmaking limitations at the time.
You're not actually a "very high skill player" your match was determined to be in the very high bracket during the time in which you were playing.... technically.
Edit: so if there were an event that got a ton of trench players playing games at the same time and limited the number of higher skill players playing then skill brackets would shift the zerg of trench players up. Let me know if this is confusing.
Right, I play in Normal, High, and Very High all on the same day. But I think he is asking about the games themselves, whether the avg game mmr falls into which bracket. There was some information about that floating around a couple years ago, but I don't even remmeber who came up with the claim.
Also, wow 98% of games are AP or AR?
associating MMR to skill brackets is a little messy, but i hope to release some numbers on that soon with a new feature launch.
I think previously someone had tested them but that was so long ago it's not really reliable anymore. Esp since we don't have a great idea of how MMR has inflated or shifted.
Regarding ranked games being more popular... I wonder how much of it is a more competitive/engaged audience and how much is the reborn account level changes. You used to need a crazy # of games to unlock ranked.
So for July/August it was only 32% ranked. It seems like there is a gradual trend towards people playing ranked, but the reborn changes made a difference. This based on match volume, not player account. So 32% of games were ranked in july/aug.
reading these numbers is so sad. i sinceerly belief that 80 % of DotA players d be better of playing LoL. And i am a 'LoL Hater'. Afterall its just a game and afterall i my 4,7k teammates complete lack skill correctly gets classified as 'very high skill' .... so soo sad.
Can you tell how many of the 11% AR games are bot games? Australia seems to be infected with farming accounts who are stuck in LPQ.
Also how many games were AR before the LPQ change (winning game in LPQ reduces LP count).
Per favore Accedi per inviare commenti.
Questions:
1. Distribution of normal, high, very high
2. Distribution of unranked and ranked matches (and custom games?)
3. Distribution of game modes (all pick, all random, single draft, ...)
4. Abandonment rate in the three skill brackets and unranked/ranked
Answers:
1. http://www.dotabuff.com/topics/2015-01-17-how-many-games-on-normalhighvhigh
75% normal, 15% high, 10% very high
2. ~61% Non-ranked, ~40% ranked (October 2015)
3. 87% AP, 11% AR (October 2015)
4.
If you guys also have question feel free to ask and I will add it. Hopefully a Dotabuff employee can shine some lights. Or if they already answered please point me towards the source so I can add it here.