Because he is a meele offlaner that is easily pushed out of the lane by a triple lane or strong duo.
And what supports and roamers below 5k dont do is following offlaner to the jungle so he has to go jungling on the otherside of the map with no access to the runes, leaving him starve for exp/gold.
Another factor is that high mmr carry-players usually prioritize winning their lanes and snowballing from there, rather than "how to counter enemy Storm Spirit". So they pick accoring to the first factor, rather than going for Anti Mage.
Meta is more prevalent in 5k I guess and they draft well, their offlaner probably fits into meta and their team comp. idk Underlord maybe doesn’t play that well in the 2%of games. I know I’m far from bluestar, but it’s what I think
5k is enough of a sample size for most heroesThen how come heroes with low pick rates tend to have winrate towards extremes more frequently?
The thing I like most about Underlord is that he has a really versatile item-build(like SF) and can easily split-push(also rat with his lvl 25 talent, 25 percent firestorm building dmg).
His win-rate probably dropped down because of people getting smarter and learning how to counter him lol.
Also this:
And what supports and roamers below 5k dont do is following offlaner to the jungle so he has to go jungling on the otherside of the map with no access to the runes, leaving him starve for exp/gold.
Then how come heroes with low pick rates tend to have winrate towards extremes more frequently?
because thats how it works?
this means u cant judge how good a hero is based on its winrate in 5k bracket across a week when it has low pickrate
nobody was talking about the weekly winrate, ud has a massive drop in 5k+ according to meta tab which is monthly stats
> this means u cant judge how good a hero is based on its winrate in 5k bracket across a week when it has low pickrate
That's not true. For example, one of the lowest pick rates in 5k is Lina (3%), and it still has ~100k games/week in very high skill.
Now calculate the confidence interval given the winrate of 45% and lets pick the extreme standard deviation of 0.5 (the real one is lower, but I can't calculate it fast enough).
The result is 0.45 ± 0.0041 for 99.9% confidence. That means that the chance that the real winrate is 46% or higher is less than 1/1000.
Several threads have mentioned this. The 5K data in Dotabuff seems to be bugged and takes into account a very small sample of games.
Моля, впишете се, за да публикувате коментар.
how comes he falls off like that?
he used to be 53% in 5k+, i dont see direct or indirect nerfs. tho he fell off af
is it because of the popularity of heroes who deal with him very well?
i'd appreciate some bluestar forum users to explain if any can
thx in advance