I just finished a game that had an early leaver. I checked my stats here afterwards, and that match is counted as a "real" match. So counting some "no stats recorded" matches probably accounts for the win/loss discrepancy.
I played only 2 bot matches, both of which did not get tracked in dotabuff.
Also I have 6 "no stats recorded" matches, 5 of which are won matches. Only one(!) of those matches was an early leave 0-0 match. Some lastet like 45 minutes. This is kinda strange. Why would those not be recorded?
The games with early leavers (when its 0-0) wont count in dota 2, but will count in dotabuff. almost sure thats the reason.
@i_Ni that is not entirely correct. Only games with an early abandon (while kills/deaths are still 0-0) will not be counted in DotA 2.
Just for the record: I've only played "real" matches so far. No diretide or w/e.
Yes, he meant when a player leaves before 5 minute count will not be recorded in dota 2
But sometimes dotabuff will record them. I don't know why but it counted a few of my games which had an early abandon while not counting a few others..
But this is simply not the case for me. In my entire match-history (according to dotabuff) there was a single game which ended before the 5 minute mark and with 0-0 kills (55919335). There is also one more game which ended within 5 minutes but not with 0-0 kills (56151630). That's one win and one loss.
Also this doesn't explain why some games which lastet up to 45 minutes are listet as "no stats recorded". Even if I take all those non recorded games into account it doesn't add up to the right number of games, wins or losses.
I will just quote myself from a couple of posts above
"Just for the record: I've only played "real" matches so far. No diretide or w/e." That includes grievling.
We don't have a guaranteed way of knowing whether a match counted or not without looking at the replay. Unfortunately, we no longer have access to the replays to determine if stats counted, so we have to take a guess. We don't include bot games and we do our best to determine if it was likely that no stats were recorded, but we just don't have a 100% way of knowing.
As I said before: this issue is older than/has been around before the new privacy option. My wins/losses were off as soon as I discovered dotabuff a few weeks ago.
I understand that as off now dotabuff does not have the ability to look into the stats deeply since there's much information no longer being shared.
I also understand that those "no stats recorded" games are most likely due to abandonments/4v5 games (which is still a bit odd, because I believe they should still be counted in DotA2 unless it's an abandonment within 5 minutes and 0-0 kills). Once again, those matches took place 1-3 months ago, so long before all the privacy issues arose and you should have had access to the replays then.
But this still doesn't account for 1 win and 7 losses in dotabuff that do not appear in DotA2. If anything this should be 5 wins and 1 loss. Also this doesn't explain some 20 games that I played with another alias apparently though I've never used another alias.
I know that people working on dotabuff have other things to do than to care about my stats. I'm not overly dependant on my stats either. Then again I'm not the only one with this issue and the idea of a public database that offers paid content is to be accurate - and we're far from 99% accuracy on my stats. I'm just curious and thought devs might want to know that there are some pending issues with the numbers.
Zaloguj się aby dodawać komentarze.
Okay, maybe this has been asked a ton but I don't feel like going through hundreds of old posts:
My wins and losses are somewhat different in DotA2 than in dotabuff. The difference is 1 win and 7 losses (both higher in dotabuff than in DotA2). Also I have played a total of 299 games (according to dotabuff) while I only have a total of 278 games with the current alias. However, I have never played any games with another alias.
The problem has been around quite some time before all those API and privacy stuff and changes.
Any explanation?