MyDotaSkill... a site that gives you a Dota 2 skill rating

#1 Posted by chiablo (924 posts) -

www.mydotaskill.com

This site has gotten quite a bit of attention, both positive and negative. It takes the last 20 games you've played and compares you with other players to determine how you rate. The result is a system similar to Starcraft 2's ladder rankings.

However, it explicitly says that if you are paired with higher tier people, you're ranking will be artificially high, and the creators have never shared the algorithm they use to determine rank, so take it with a huge grain of salt.

Here are the levels:

Beginner < 20

Copper > 20

Bronze > 100

Silver > 200

Gold > 400

Platinum > 600

Emerald > 800

Pro > 1000

Jadeite > 1300

Mithril > 1500

It does take about an hour for it to generate, and it's easiest if you already have a DotaBuff profile (which everyone should).

#2 Edited by EXTomar (4690 posts) -

No no and no. This thing is borked because it does little to track skill. Basically because the way Dota 2 works, 20 games isn't enough to determine a trend of anything let alone rate your performance. And unlike SC2, Dota 2 has 9 other actors building and interfering with your performance.

Personally I think these things are toxic. I have seen what "services" like this do in other games where people are screaming before the match they got matched with a loser even before the match is started. If they win they think it is because of their awesome high skill. If they lose it is because the other guy who has a terrible score.

#3 Posted by chiablo (924 posts) -

@extomar said:

No no and no. This thing is borked because it does little to track skill. Basically because the way Dota 2 works, 20 games isn't enough to determine a trend of anything let alone rate your performance. And unlike SC2, Dota 2 has 10 other actors building and interfering with your performance.

Personally I think these things are toxic. I have seen what "services" like this do in other games where people are screaming before the match they got matched with a loser even before the match is started. If they win they think it is because of their awesome high skill. If they lose it is because the other guy who has a terrible score.

But something is better than nothing. Right now in-game, the only thing you can use to judge where you stand is game time, level, and win rate. All of these are fairly meaningless when it comes to tracking progression. I agree that there should be some kind of graph based on your entire history so you can see trends and also account for inaccuracy due to being in groups with higher skilled people.

And apparently, Brad is pretty good. :p

#4 Edited by EXTomar (4690 posts) -

On the contrary, "nothing" is better than a made up number that people think falsely informs of their status.

Part of the big issue is that you don't need to be a statistician to figure out what is wrong here: They are trying to figure out how closely you perform against top tier players from the last 20 games. That is kind of fanciful where unless you are also a top tier player playing in a top tier tier team playing in CD/CM, it is a meaningless comparison because all of the factors measured are wildly variant. You could be close to performing like NaVi.Dendi for one game. Was it because you were suddenly improved that much? I highly doubt it. Was it really that your team doing really well? Or was it the other team doing really poorly and someone wanted to extended it the match an extra 10 minutes to get more gold/exp and another Rosh kill and along with merciless pick offs of the other team because someone wants TO PAD OUT THEIR SCORES FOR SITES LIKE THIS? The next game you aren't even close so what does that mean???

I have never, ever understood this drive from people to have some sort of rank or number or whatever in games like Dota 2. I do think there is need for extend stat tracking that was available in Dota but also the stat stuff found in CS:GO and TF2 that are personal and are only meaningful for you on that specific machine. But a ranking that can be looked up on an external site??? Why?! Dear god why!?!?

#5 Posted by TowerSixteen (542 posts) -

@chiablo: Gotta agree with EXTomar, these things are bad news, divisive, poorly reflecting reality, and just generally a matchstick thrown into the big'ol pile of tinder that is that community.

#6 Edited by SirOptimusPrime (1997 posts) -

@towersixteen said:

@chiablo: Gotta agree with EXTomar, these things are bad news, divisive, poorly reflecting reality, and just generally a matchstick thrown into the big'ol pile of tinder that is that community.

Dotabuff already generates a nightmare as it is. This is just begging for horrors.

@chiablo said:

And apparently, Brad is pretty good. :p

This is how I know this is both bad for the community and straight up bad.

P.S. Love you forever Brad.

#7 Posted by Sackmanjones (4689 posts) -

I haven't played dota in just about a month. I feel freeeeeeeeee

#8 Posted by BabyChooChoo (4424 posts) -

Personally, I also never understood why people need a visible rating in games like this. There is literally always room to improve and if you can't recognize your mistakes, send your replay to a friend or post it on Reddit, PlayDotA, or, hell, even here. Let someone else critique you. All these rankings and things just seem so insanely arbitrary especially when you take into account the sheer amount of possibilities present in any given game.

#9 Posted by Akyho (1602 posts) -

@extomar: I agree.

I've had matches newly playing Ursa with a group of GB duders, the lanes were falling it was looking bad. Then once I had felt I jungled enough I came out and ate all their faces, the team hurrahed and we won with a landslide in the end thanks to me playin as Ursa.

Do go "WOOT ME I AM AWSOME I AM PRO!" I go "That Ursa pick came through, wow Ursa is powerful. The team tactics of letting Ursa jungle and the folks holding the lanes, allowed it to pay it off. Glad they did not have much stun or else It would have been a tougher fight."

So yeah many many different things that lead to you winning 10 games in a row or losing 10 games in a row. Russians being a major point on what happens.

#10 Posted by Cronus42 (275 posts) -

Can confirm this thing is broken. The way it measures "skill" actually skews it towards supports. Also like OP said it is very easy to get carried into insane skill numbers, regardless of your actual contribution. There are too many other factors that this thing can't measure to account for skill. My own rating is mid 300's, same as the friends I play with. Based on their scoring seems low, but when you could still easliy check brackets my friends and I were almost always very high. So unless we've all gotten a lot worse, the scoring is really heavily weighted towards one end.

Yea it's broken. Seems like a good idea but it's way too arbitrary, way too gameable, and honestly could prove TERRIBLE for the community.

#11 Posted by pyromagnestir (4305 posts) -

Finally we can give up on those arcaic and meaningless stats like KDR and win/loss ratio and the era of Dotametrics can begin! People will now be able to understand when I start talking about:

  • DORP = Dota over replacement player
  • Win shares
  • ULR - ultimate lane rating

and more!

#12 Posted by Slag (4274 posts) -

@cronus42 said:

.

Yea it's broken. Seems like a good idea but it's way too arbitrary, way too gameable, and honestly could prove TERRIBLE for the community.

do you really think the site is going to get that popular among the player base?

I'd be surprised if even 5-10% of players know about it...

#13 Posted by ryanwhom (290 posts) -

these seems like something the worst kind of people in Dota would be all about

#14 Posted by UlquioKani (1046 posts) -

I got 897 which is pretty good apparently but I don't think 20 games is enough data to determine a rating. I would love to know how they calculate this though.

#15 Posted by Nadril (527 posts) -

It seems pretty crappy to me from what I've seen so far.

The problem I have with it is there is no way to tell what level a player plays at. Sure, you can try and guess at it based off of how other players do in the game. The problem with that, of course, is that those players are also playing against better players.

Like sure, I could go into a low skill game and absolutely fucking dominate. My team would also do well since, by proxy, they would be getting assists and such from my skill. That doesn't mean that I'm a good player just because I and my team did well.


It also just takes statistics (IIRC) from your past 20 or so games. That really is not a reliable enough sample to actually figure out anything.

I'd really like for Valve to just come out with their own rating system so we'd stop getting these kind of shitty versions of them.

#16 Posted by Slag (4274 posts) -

So this just relaunched the other day. I'm not sure it's a good thing, but hey it's there with lots of new graphs and whatnot.

Seems to have completely revamped the scoring system as well. I'm not really sure how they are computing some of the components though, (especially nebulous ones like teamplay).

This edit will also create new pages on Giant Bomb for:

Beware, you are proposing to add brand new pages to the wiki along with your edits. Make sure this is what you intended. This will likely increase the time it takes for your changes to go live.

Comment and Save

Until you earn 1000 points all your submissions need to be vetted by other Giant Bomb users. This process takes no more than a few hours and we'll send you an email once approved.