as sano already said, the top 50 needs to be reworked.
It would be very easy to exploit a top 50 system that does not take MMR into account. Unfortunately a system like that will be less indicative of skill than the one we have now.
Well, we have 2 options regarding top50
1) Keep the fail algorithm that uses DBR which doesn't exist anymore and wait (and pray) for an official ladder system
2) Make a new one without DBR - won't be as good as one with MMR, but we can't do anything if Valve doesn't share this information
There really is nothing dotabuff can do to gauge the top players anymore (even before it wasnt all that accurate)
major things that could kill your idea are
-best hero does not always = most played hero (so overall win rate will hurt people who play heroes they suck with)
-bad players sometimes deny and last hit instead of taking advantage of key objectives (ex. bounty hunter making battle fury and farming)
-In reality the only way to gauge the performance of a player is to base it how they perform with the same teammates. (ex. I play bounty hunter really offensively and even if my teammate are overall good players if they play extremely passively we will not work well together. I could be top 10 offensive bounty hunter players while below the 5000th top defensive bounty hunter players)
I just want to say I've read this thread and sano's thread. We appreciate feedback and suggestions, it helps us improve the site.
I have seen safe base post maybe 6 times. I have seen another staff post once. I am sure he posts a lot more than 6, I don't read every thread.
댓글을 작성하시려면 로그인이 필요합니다.
DBR is no longer available and the current top 50 is no longer viable or updated.
Can GPM and deny be used to somewhat gauge skill bracket? I'm sure low bracket will be removed.
We have stats for K, D, A, LH, DN, XPM, GPM, HD, TD, hero win rate and overall win rate also. I am sure there is some way to make a new algorithm for a new top 50.