Two months ago I wrote an article discussing the trends of pub Anti-Mage, specifically that his win rate is in a very steady downward spiral. Two months ago AM’s win rate was hanging around 41.6%. For the first half of July, his pub win rate has dropped again, this time to 40.6%. From the data I have available to me, 40.6% is about the lowest win rate AM has had in the last 2 years. If may be off a few 0.1%, but it’s essentially his all time low. His win-rate is still dropping, a very clear downward trend. But why is this the case?
Itemization is a big part of a hero, so I’m going to compare the changes in build rate from 6.80 to 6.81 (and 6.81b) in order to see how people are building AM differently in the last 2 months. He’s dropped 1 raw percent, and itemization is going to show us some valuable information. Keep in mind these are items Anti-Mage finishes the game with.
The same trend is observed as last time I ran the numbers, that Battle Fury is constantly being built more and more. This time the other items don’t move too much, whereas Battle Fury moves more than twice as anything else. Battle Fury had a 40.5% build rate in 6.76, and it increases every patch. Although I did use a different data source last time causing some discrepancy in the exact numbers, the trend is still the same.
We now have two trends: Battle Fury’s build rate on AM is increasing over time, and AM’s win rate is dropping over time. This doesn’t bode very well considering that a large percent of the Dota population are very firmly of the idea that Battle Fury is the way to build the hero. There’s a couple of routes we can go with this knowledge.
Correlation and Causation
The first is the knee-jerk reaction to say that just because the two trends are happening at the same time, doesn’t mean one is causing the other. Which is entirely logical, so lets add in 2 more potential causes: changes in Anti-Mage’s pick rate, and changes in all heroes pick rate.
I record and analyze the difference in heroes pick rates on a month to month basis, so I have that data directly available to me. My post on last month’s changes (June), can be viewed here. Anti-Mage increased 0.66 raw percent in June, and decreased 0.22 raw percent in May. For comparison a notable change in May was that Ember Spirit went down 7.22 raw percent. So basically Anti-Mage’s pick rate goes up and down slightly each month. Thus far this month it is up about 0.28 percent, but is starting to go down again.
I don’t see anything too drastic here, some heroes get massive surges in pick-rates, but Anti-Mage is not one of these heroes. So what about a surge in other pick rates?
Increase In Counters
Tinker and PA have both surged in pick rate recently; AM is good against Tinker, but bad against PA. So I was going to run the pick rate numbers on all 107 heroes, and then see if the average hero now counters Anti-Mage more than before, but I realized I don’t have to. You see, there’s a very basic Dota principle that I think some people seem to forget. Specifically, if the enemy has heroes that your hero matches up badly against, then you should build your hero very differently. Anti-Mage has a terrible 32% win rate against Bloodseeker, which is well below what it should be. If you’re an Anti-Mage against a Bloodseeker, then you might want to go items that allow you to survive his early ganks.
Even if people are playing Anti-Mage counters more than before, it’s technically somewhat irrelevant. Because if Anti-Mage is getting countered more now than before, why is he building more Battle Furies? Why would a hero who is getting countered more, be going greedier builds? You would think that if a hero is getting countered more, the most logical thing to do would be to build safer, not rushing more of a farming item that gives virtually no survivability. Which is likely to result in you dying over and over again to their strong early + mid game presence.
Which leads me to thinking that the true cause of Anti-Mage’s declining win-rate is player’s overwhelming stubbornness. That is, people are becoming more and more convinced that there is only one way to build the hero, which explains the constant rise in Battle Furies. And even if they are heavily countered, they still build in the same greedy way. I assume they go “if I wanted a carry that came online earlier, I wouldn’t have picked AM“. Since Anti-Mage is going worse, they instead should be going “I should build in a way that makes me come online earlier, because of the way the game is going” even though it’s not what they originally intended to do.
Time and Place
Personally, I feel like the meta is moving away from having heroes that AFK farm for long periods of time. The meta (as well as players) seem to punish players that attempt to abandon their team in order to farm more than every before. Even if they attempt to get other trades, their lack of ability to join team fights in the early and mid game is devastating for their team. And since pubs generally don’t have a well composed team, going full out farm mode is likely to cause crossroads with the other players on your team. Maybe without them even realizing, instead blaming it on the hero, instead of the way you chose to play the hero.
Which leaves me with very simple advice for Anti-Mages. Yes, Battle Fury on Anti-Mage speeds up your farm incredibly. But by building that item you are committing your entire team to a playstyle that they are probably not suited for. This is not a pro-game; where everyone is on the same page, with a cohesive game plan and strategy. (Not to mention there are very strict timing windows for these sorts of items.) This is a pub, and the stats seem to show that going greedy is not working out for Anti-Mages. Whether it be because going greedy against heroes that counter AM more than before is a bad idea, or just because doing so in a pub is bad idea. It doesn’t matter, the end result of a declining win rate is the same. It is being stubborn, and resisting the natural change the game has over time.
I’ll keep my eye on his trends, and post again in a few months. But in the mean time, maybe rethink what you know about the hero? You should think of one thing every time you see an Anti-Mage pick, including when you pick it. That is 40.6%. His win rate from my sample is exactly what you should be thinking. 40.6 per cent; the magic number.