Great article, as always.

However, I would recommend changing the phrase 'something that really doesn't fit into my head'; ' something that really doesn't fit in my head' being used in pretty quick succession (Alch and LD) to avoid the Deja Vu effect while reading these.

Keep the good work Dotabuff!

good article. (second)

hey guys

Well Faceless Void isn't dead yet but he might still be forgotten in low level pubs since hes not the shit now.

I think the fact is that people in Pubs can never get no where close to Pro matches GPM if they're not flat out winning. Pubs I'd say sub 3500 MMR or so consistently get items a good 5-15 minutes later then they should get them compared to PROs. And most of the games played are well not 4000+ MMR matches. In Pro matches the Hard Carry has all it's items by minute 40. Naga, Spectere, Anti-Mage and all other "flash farm" heroes while in pubs you see Naga Radiance at 25 minutes and at 56 she still sits at 5 slots maybe the 6th one is like a Ring of Aquila.

We saw in the EG vs C9 match that people actually kept up with Naga Siren Radiance farm while not losing the game. That usually never happen in Pubs as people go fight each other vs farming passively for some sort of timing push. Pubs having 120 kills is like every other game while Pro matches having 100 kills is every 10th match.

So I think Lone Druid, Alchemist, Anti-Mage, Naga, Morphling and all the other good heroes that are in the sub 50% tier are there because people simply don't farm well enough. I speak from experience here, looking at blacks recent games I can't even hope to get that amount of farm on anyone.

I used to hate on Lone Druid for some reason but I started playing him this patch and having mana on Spirit Bear is a huge buff. Dotabuff pub win rates have never reflected the real strength of any hero (Earth Spirit is the best example). I generally think that the win rates of heroes with heals and defensive abilities are skewed so high because uncoordinated teams don't focus enemies in fights, and deciding as a team to engage/disengage is clumsy when you aren't communicating as a team.

I'm really surprised to see Timbersaw sitting at a sub-50% win rate. The new aghs upgrade is insanely powerful.

Surely the fall in Chen's win rate has a lot to do with inexperienced Chen players wanting to try controlling ancient creeps?

I don't really get this part about Void: "he did not necessarily gain as much in terms of assist/teamfight gold"
Can someone enlighten me about the mentioned assist/teamfight gold?

While it is a nice blog. I'm wondering why there is a Naruto Opening subtly linked?

@big black cock, because of the changes to gold earnings per kill (comeback gold) heroes that can deal aoe damage and get lots of assists receive lots of gold per teamfight

Faceless Void is a single target hero, therefore does not completely benefit from this change whereas aoe monsters like Medusa and Spectre do.

Thanks, you got my name wrong tho (hehe).
I don't really remember but before this patch you could have an assist just by standing near a fight while now you have to deal damage to the hero/s?

since when is DP and Void an average stomper!?!?!?

Excellent post. Early game isn't dead, but, I think, more proportionately risky compared to picking a strong mid to late game team. Meaning: it takes better execution to pull off an early game strat than pre-6.82. Also I love the respect given to Brood Mother in your post.

the armlet nerf was what killed lone druid. nothing else can bring him back

@swag since the win rate said it was? he doesnt rank them. It is based on win %'s.

in the chen tabs ya decreased ofc cause not anyone can play well microheroes and microing ancient is a lot harder than controlling a meepo

[you dont wanna give away free gold to enemy]

p.o.
-smurf-

I think the reason why LD's winrate is that low is because he's a high skill ceiling hero and, even though he has 12 slots available, his lategame is not that good to be honest. Even in this greedier patch you will not see a 12 slotted LD very often anyway.

Fun fact: Razor's win rate was shit at TI4 as well. He is overpicked as a first draft. I consider him a great counterpick only.

razor is a great counterpick to heros like troll warlord and faceless void since both of these heros deal huge amounts of right clicking damage. Razor can just suck their damage and rape them provided the player knows how to play razor and razor isnt underfamred.

I think it's kinda strange how Lich's winrate dropped when he's been receiving nothing but buffs for the last 5 patches, while Wraith king's seemed to increase when he received substantial nerfs. Neither are hard heroes to play. Perhaps because Lich is an early game support while Wraith king's armor nerf helps him actually die faster

Wish ratings like those didn't exist.
They serve no practical purpose (anyone can see winrates already), and all they do is make good but underrated heroes even less popular and overrated heroes even MORE overrated, which is completely undeserved in both cases.

Worse yet, now when someone picks a so-called "low tier" hero, the entire team will flame and report him no matter how well he does, and blame it only on him if they lose.
It is truly a shame that even the gems of the Dota community, the statsmen and analysts, end up working against the game in such an inglorious way.

I honestly dont understand why people repick lonedruid , io , enchan , alchemist. These 4 heroes are so fun to play especially for lonedruid , u get to build up to 12 slots ~.~ and alch overs a fast GPM since in pubs u wouldnt take skill 2 until lv 10 (im assuming for solo players who doesnt intend to kill)

a graph showing change in win rate would be really neat for patch analysis.

Terrorblade is not a pub loser, he is a pub stomper, BUT ONLY if played correctly, which requires above average microing, map awareness and game sense. The hero is really powerful, and win-rates in pubs aren't always 100% accurate. I've played 8 games so far with Terrorblade (check them out on my profile) and lost only two. Average GPM with the hero: 680. Average XPM: 618. KDA: 2.89.

And this is by jungling him, every game. I'm around 3,3k MMR if that's relevant. But even at higher tiers, this hero is absolute bonkers. There's a reason he's banned in the comp scene a lot as of recently. He farms better than a Naga Siren without Radiance (Radiance is a game losing item on him for the record), and utilizes the farm much better. Radiance Naga doesn't really do anything except split pushing until she gets Heart + 1 item, and that comes after BoT, Radiance and Manta. Whereas Terrorblade wrecks a whole team with just level 8, basic boots and a Yasha. He also pushes, way better than a Naga ever will. The only issue I see on him as of now is that you're reliant on Metamorphosis to take objectives early or teamfighting.

Eh, a lot of the people lower in the list seem to only be there due to lack of skill. Chen, Earth spirit and Io are probably lower because people can't play them well unlike simpler heros like Wraith king or Slardar. Seeing as these are based off of percentages, I guess this is expected.

People tend to throw as brood in the 4k+ mmr bracket because they get cocky and overconfident, they take down a ton of towers and get a bunch of solo kills in the offlane and think they can buy stop items, then they suddenly drop off heavily in the late game and continuously throw themselves at the T3 tower by themselves and feeding all that advantage back to the enemy team in this patch. The last game I played against a brood we were losing horribly but the brood was being a tard and purchased an ethereal blade and I think started building a dagon or something... Was completely useless in the lategame and we made an easy comeback since brood just went from being our biggest problem to an easy pick off.

I don't agree about the legion commander. I believe she has the ability to play at any position and scales pretty heavily in late game

I think Alchemist is just a bad hero, the trenchy way to run him is to max greevils and get a midas so he just hyper farms, but he's so easy to gank and is so squishy (even with his massive regen from his ult).

If you run him as a support / initiator then you have to rely on his stun hitting multiple people and it's just too unreliable without a blink or shadow blade, and on top of that he can't contribute much to fights in the early - mid game because he's (as I said) too squishy.

I think Acid Spray is SEVERELY overrated, it's very mediocre anti-push and most carries can just tank it / fight under it without even caring. It will fall off very quickly and only really be good as an ancient farming tool.

The benefit is supposed to be he can transition into a carry and get items up faster thanks to greevils, but his early - mid game presence is just so poor unless he's able to get a lot of successful ganks off with concoction.

Alchemist should be picked much more since u can literally get 1000 gold per minute when u maxed out your greed.
And believe me it´s only worth to pick up a divine when u got a blademail a heart of tarrasque and a bkb.

This is so yesterday.

blog post on 6.82b comes out at the same time as 6.82c. nice post just it would have been better a bit sooner.

it comes out with the end of a patch to see how the patch has affected hero winrates

its still viable dude , the only thing change in 6.82 c is in here
http://store.steampowered.com/news/14675/
u have brain , use it , u can predict what happen in 6.82 c base from changelog from 6.82b -> 6.82c

its a data dude , even data ti 4 only can get after match
how come u make a data 6.82 c when 6.82 c launch tomorrow for example
maybe we need a paranormal that can predict future (and we no need brain again)

Nice one. Want to see more articles like this.

The problem with Ench and Chen on pub is basically almost no one ever pick/played an active support and most of people who played support were people who were forced to play support because the other 4 were picking carries. Most of people who pick support rarely playing an active kind of support, they only tried to zone out offlane or worse, trying to steal CS

Ench and Chen on competitive wasn't picked that often too because of different but similar reason, if Ench/Chen didn't made any meaningful impact early = game probably already lost

very well done article, but dayum son it hurts to see my beloved heroes being so low on this list(lone druid,lycan,nightstalker,naga,AM),the only one doing well is good old undying

I'd say that the Phantom Assassin win rate increase is due to the nerf to her counters: LC, Void and Doom no longer disable her passives. It's funny to see a void that wastes his ulti to deal almost no damage to her. Plus, the change to blur is kinda tricky because your enemies can forget about PA since they don't see her on the minimap. As a PA player, these days i feel like i'm getting ganked a lot less mid game.

If you played Lone Druid in a really late game scenario it'd be quite obvious why his winrate dropped, despite the fact that he's a 12 slot hero, the bear is extremely easy to focus down in the really late game by a farmed carry, and now that most carries do get farmed he became quite weak. I'm a little confused as to why Lone Druid wasn't as popular in the last patch, as the hero is strongest in the early mid game with 5 man pushing with a radiance, at a time where no one will ever kill the 2700HP bear which deals tonnes of damage.

Alchemist's winrate is explained by Icefrog's recent crusade to eliminate all heroes with a bad early game. When Void, Spectre, Morphling, and other super hard carries had bad early game to compensate for their late game dominance, Alchemist was a hero who was the reverse of this. A good early game carry/ganker who wins early by accruing a good gold advantage early.

Both of these changed. Hard carries have it easier then they ever had before in 6.82 and gold gains have shifted tremendously so that these heroes can stay on pace with Alchemist while being a harder carry. Also since games drag out, Alchemist's early game advantage is stretched even thinner, leading to his abysmal 40% win rate.

I do wonder what would happen if you split it between support Alchemist and carry Alchemist. You'll probably find the former has a larger win percentage than the latter.

I understand what you're doing here, but some of these 40% winrate heroes are better than you think...
If you want to see my ember spirit, I'll show you how 40% is more like 80%- Same with Earth and Timbersaw!
I doubt you will do so but I think you should probably note that some heroes have huge winrates when you know how to play them- Meepo for example- When I get a good Meepo in one of my matches, rather than a randomer... Once again 40% is more like 80%.
I suppose It could be argued that this can be said for the other heroes too, personally, however, I'm not the best CM or Spectre and I reckon someone who is good at Spectre will still be far off my Ember Spirit (Not to sound cocky but I can play that hero!)

TL;DR- Some lower winrate heroes are just hard to play but should have like 60% winrates when you play them properly!

>50% - hard to kill these heroes, easy for playing.
<50% - hard to playing, easy to kill if on hero plays noob.

2.5k player article
you should understand that winrate doesnt mean how good hero is :)
Tb Meepo Lycan pub losers?
U WOT M8

(Y)eeeeeesssss

To be honest, I'm a Lone Druid player and I've played him enough to tell the reason to the fact that the winrate is so low.
Basically, it is a micro intensive hero that you need to control using some Group Control keys or scripts (if you are tryharding), and most people just dont use it. I've already seen some people just using their mouse to select them both and thats how the entire game goes along.
Second of all, he is situational these days. Heroes like Viper, Drow Ranger, Phantom Lancer, Ursa and many more can counter you really hard, putting out immense ammount of damage towards your bear and fucking your fight out. That also explains the Winrate, as these heroes listed are pub stomps hero and you can find them and litterly every 3rd game out there.
And at last, the new item builds are shitty - MOM is a bad item pick on him, as he is badly going to the mid-late game with it, going well only for the early, as the bear wont take that much of dmg with MoM on in min 35 for example. Also, people are rushing Radiance and getting it just about too late sometimes (min 25), and that just destroyed the power of LD in those situation. Also, there are just some hard carries who out-carries you, like Medusa, PA, PL, Void, Sven and Terrorblade.
I believe that he will gain his power once they will let him gain stats again, making him viable for late game again, becuas 2700hp is just not enough, and you cannot buy a good slow for late game on him without regreting on the waste of money when you buy that Eye Of Skadi.

I still have a good WL with him, becuase I run my own, unexpected (for many players seeing a Lone Druid with Radiance, going like "wtf is that?") build and tactic with him, but still, he is out-carried in many games becuase of the usual pub-stompers just scale better than the bear.

The reason for the drop in win rate in pretty much all carries except for the ridiculous spectre is that people just can't farm properly. I myself have witnessed hundreds of carries miss 50% of their last hits in the first 10 minutes, while constantly getting caught out by heroes like pudge, bloodseeker etc. Spectre is the odd one out because she is just flat out too easy to play for the reward it gives. No other carry matches her lategame, not even medusa. With this in mind, it seems broken that she can easily dominate any pub after the 15 minute mark if given an easy lane. This is even more true en high level games where a spectre can easily achieve a 15 minute radiance timing with no deaths. Considering how hard it is to crush her in the laning stage because of how tanky she is, this just feels flat out broken. From this point on you just farm jungle, and then ult when the fight starts somewhere. All enemy heroes will melt from radiance as well as her desolate abilty. The only counter to this hero is to either crush her in the early game, (her timing must be slowed by at least 10 mins for this to be effective) or to simply go for a 5 man bkb strategy with a ton of lockdown, seeing as spectre really likes to avoid building bkb if possible.

This is why you will see heroes like spectre dominate, whereas squishier, skill heavy carries such as anti-mage, terrorblade, gyrocopter, lone druid, etc. will be crushed in the bulk of their games.

lycan high skill ceiling lol?

As someone who almost solely plays NP, I'm surprised this article didn't mention the huge drop-off in his win rate. I can say from experience that it's due to the changes in two very popular heroes: 1) Riki invis starting at lvl 1 messing up his farming game if he jungles, and 2) PA's blur rework really, REALLY fucks up your map awareness.

PA is a ridiculously hard counter to NP now (not that she wasn't before already), and I think considering how popular these two heroes are, that PA's win rate has increased dramatically and NP's win rate has decreased by nearly the same amount...I don't think this is coincidence. Just about all the stats corroborate this line of thinking.

PA used to be easily countered by legion (perhaps with blademail), but now legion's ult doesn't disable passives, so PA dodges quite a bit. I'm assuming thats part of PA's increased winrate.

I think pubs just don't know how to play LD. It's really a shame considering that it's my best and most played heroes. Altho my MMR of 4.1k isn't considered semi-pro or pro level, my LD has a win % of close to 60%. I definately think that it's just the pubs who don't know how to play the hero :/