Hey gamers, i’m back with another “I spent 300h” manually doing something cause I can’t program and also didn’t get called to cast ESL qualifiers.
Basically, I manually looked through dota2protrackers games that are 60min+ long and also high MMR (avg 9k+) and I crunched through the data to find some cool stats about neutral items:
I’m going to start with tier 5 items because this is the hardest to get get good data on and the one people play the least:
Couple of things to take away from it:
- Ex machina is by far the most overated tier 5 right now. Even though the ceiling of it seems insane the execution is super hard (you can watch the video at the end if you want to understand what i mean), the mana cost is prohibitive even for heroes with a lot of mana and its also not a straightforward item
- I wouldn’t really say book of the dead and apex are as broken as they look because the number of data points is quite small but it does seem to suggest they are at least underated. At the same time pirate hat is definitely up there as one of the best ones. Very low bar to use and huge upside every time
- Items that have passive effects are in general really good (seer stone, sty desolator, mirror shield, giant’s ring). They are mostly passive and even in the highest of mmrs it seems on average a safer bet than complicated items that have actives (book of shadows, ex machina, arcanist, etc)
Now let’s take a look on pick rate/winrate for each tier of neutral items. This is obviously a bit skewed since I selected games that went long we might not have the best data possible but its still interesting to look at it specially since no website really gives you the data to look at it at this degree.
TIER 1
- The first thing I noticed is tumbler’s toy winrate being the lowest one! I think most people still talk about it as if the item is literally the best one avaiable and at the same time, the stats don’t seem to agree specially since in this case it is the 3rd most popular one so lack of data isn’t necessarily an issue here
- Another big thing that jumps to mind is how occult bracelet is the 2nd highest winrate despite the gigantic ammount of nerfs it suffered in the recent patches.
- I’m not sure if I am biased (could be) but once again passive items seem to consistently have higher winrates then the active counterparts the only outlier being trusty shovel.
My initial idea with this post was to go over every single tier of neutral items but I think this would be way too boring and no one would read it so yeah idk, if you guys think this is cool I can post more updates in the future. I am still going through the data so there’s a lot to discover. In case you want to watch the video I made about the data set you can check it here!
Sorry, but looking at winrates and then making a post that says Everyone/Pros suck at tier 5 items is just ignorant. You’re leaving out so many variables.
If team A has 3 heroes that have many active items and pick 3 ex machinas, and team B only picks 1, and then team B wins, WAOW DAE EX MACHINA SHIT? What if the problem was in the draft, or maybe the players didn’t even get to use it because they died in the fight too soon, or maybe the players just didn’t utilize the item to its potential (which you allude to yourself with the skill ceiling) or maybe it was legitimately the best shitty choice out of the 5 options they got.
What about the fact that some heroes will always pick some items and not others, or the fact that you can’t have all the items in all games.
Not to mention data skewing via “we’re already winning, so I can take Apex/Pirate Hat and just have fun instead of tryharding with ex machina.”