[Megathread] DotA 2

[Kapaya]

Thing is Underskilled, I can't see this hero dying if played correctly. Beyond the ridicilous armor he gets with each hit, which whill make harassing him a pain, that 4 second hook makes it fairly easy on him to escape (while it could also screw him up badly if placed wrong!). I can see a Shivas + Hex on this guy is gonna make heroes go splat when he pops his ulti. Hell even Necromicon might be something to go for? Gonna be interesting to see how he actually is!

[Lysah]

That's... depressing.

I disagree. The system seems to recognize that if you only win when you queue with 3 of your friends you do not deserve to be in ultra high MMR. Likewise, it seems to recognize that if you only lose when you queue with 3 of your friends you do deserve it. I'm not sure how it measures anything, but I do notice that my MMR does not get punished much for losses. If I lose a bunch of *solo queue* games in a row, my MMR might start changing a bit, but it usually jumps right back up after a few wins in a row.

At best, I would believe that it keeps track of two different MMRs for each player - solo and not solo. I kind of like knowing that when I lose with my bad friends that I'll still be recognized as a good player by the system. Furthermore, plenty of pubs love to look up enemies mid game and laugh at their win rate, information like this proves that win rate is a largely irrelevant if not an altogether pointless indicator of skill.

[Underskilled]

Thing is Underskilled, I can't see this hero dying if played correctly. Beyond the ridicilous armor he gets with each hit, which whill make harassing him a pain, that 4 second hook makes it fairly easy on him to escape (while it could also screw him up badly if placed wrong!). I can see a Shivas + Hex on this guy is gonna make heroes go splat when he pops his ulti. Hell even Necromicon might be something to go for? Gonna be interesting to see how he actually is!

Defensively, yes, he sounds like a powerhouse. However, the reliance on trees will pose some awkward situations. In base, for example, he'll be VERY weak, which will make him an odd hero late game. Mid will also pose an awkward place for him as well.

On the other hand, it will be interesting to see what sort of stuff he can pull off with nature's prophet.

I disagree. The system seems to recognize that if you only win when you queue with 3 of your friends you do not deserve to be in ultra high MMR. Likewise, it seems to recognize that if you only lose when you queue with 3 of your friends you do deserve it. I'm not sure how it measures anything, but I do notice that my MMR does not get punished much for losses. If I lose a bunch of *solo queue* games in a row, my MMR might start changing a bit, but it usually jumps right back up after a few wins in a row.

At best, I would believe that it keeps track of two different MMRs for each player - solo and not solo. I kind of like knowing that when I lose with my bad friends that I'll still be recognized as a good player by the system. Furthermore, plenty of pubs love to look up enemies mid game and laugh at their win rate, information like this proves that win rate is a largely irrelevant if not an altogether pointless indicator of skill.

That sounds pretty fair to me, actually. I only ever play duo-queuing with a friend of mine, and I'm not afraid to admit neither of us are really that good, but we've got a massive advantage over most lanes, simply because we can communicate. If it ranked us solely by the wins and scores and such, we'd be up against players way out of our league.

[Lysah]

Well, people who queue with me are pretty much guaranteed to be placed way out of their league, but that's just a result of the system trying to be as fair as possible. I will admit that it's pretty much impossible to balance a game with two players on opposite ends of the MMR spectrum - either the lower player feeds hard or the higher player automatically wins. Still, it's pretty annoying and makes no one want to play with me, which is why I support a first come/first serve no-rating league.

I believe you do still take an MMR penalty for queueing with friends. A direct result of this is the page 1 games are almost always teams of 5 against each other. It's not because these are the highest MMR players in the game at the time, it is simply because they have MMR added to them when the queue as a team. A two man team probably isn't that noticeable though, since you can only really win one lane out of three with that kind of "stacking."

I think Valve has a pretty fair system going, but the quirks of it can be extremely frustrating to people. One direct problem is that a zero sum system that tries to keep people neutral is virtually guaranteed to give you allies you cannot win with if you win too much, rather than just bumping up your MMR until you are the reason you are losing because you are out of your skill bracket. I'm sure everyone in this thread can remember a time where they won 8 games in a row and then got a jungle Juggernaut ally who went 0-6 against neutrals. Sample size, perhaps, but it's still frustrating and easy to blame the system.

[Hermanni]

I agree with it

It's understandable, but while your experiences reflect mine rather well I still haven't found any evidence for MMR to be based on anything else. I figured it might very well be possible that when your MMR is far below or above the average it wouldn't be affected at all. The rating worked like that in HoN, if I solo queued at 1900 rating the game told me I'd be +5/-5 depending on if I won or lost in an evenly matched game, but if I queued with friends in 1600's the game would give me a +0/-0. If that were true, and I'm assuming it is, then it would make winrate a potentially very skewed metric for approximating MMR.

I'd consider it possible that the solo/team MMR's are separate, like you suggested, but I imagine that would mean my team rating would have to be pretty bad, which doesn't seem to be true either.

I had awful winrates queuing with some friends (ranging from 20-40% over a lot of games) which even made my winrate drop below 50% at some point (yet I never stopped regularly being in page 1 games), but after I made another account it got better. If you have bad conscience about making games too hard for your friends then I can only suggest you make another account too (I imagine everyone has enough invites) unless making games too hard for whoever you play against isn't more or equally taxing for your conscience.

[Lysah]

I had awful winrates queuing with some friends (ranging from 20-40% over a lot of games) which even made my winrate drop below 50% at some point (yet I never stopped regularly being in page 1 games), but after I made another account it got better.

Yep, same for me. I always laugh when I look up people I know on dotabuff and look at their most played with list. They have a bunch of friends that they have 80% wins with and I look at mine and every single friend I play with is negative. Must be nice to win all the time when you queue with friends =p

[JimPaladin]

http://pastebin.com/Q4E3HDCE

It is time.

[Kuntantee]

Allow me to step in here and just remind you that the arguments you have provided in the last week or so have yet to be backed up by any real statistic. Just because you have an argument does not mean that it is even remotely accurate. Arguments regarding what heroes are over powered/too strong only backed up with hyperbole and anecdotal evidence mean (contrary to popular belief) fuck all.

Furthermore, not complaining does not make you (or anyone else) any more rational or constructive.

Straight up: You think something is too strong? Too weak? Too one-sided? Broken beyond repair? State this. Back this up with statistics supporting the claim you just made. Bam, rational argument. Statistics are super easy to fetch, the Dota 2 API is incredible and there are multiple parsers for both professional games and public games.

My sincere apologies if you did in fact provide facts in any of your previous arguments, it may very well be a case of selective memory.

Actually you are right. I did not provide any statistics and I stated that I won't because mathematically speaking they are bunch of bullcrap. Statistics only applicable if there are no variable other than the subject but that's not the case. So coming here, throwing arguments based on some pseudo statistics is not a better way to make an argument storng to me because I do understand mathematics(not that you need some sort of deep understanding to find those statistics are useless). I love theory-crafting in DotA and if something makes sense when some1 writes here than that's a good enough argument on my side.

BAM! Statistical theory proved you wrong!


@Hermanni
I thought you are complaining about arguments again.

@PizzaSHARk
If you are expecting poisons, they are very easy to avoid. If your "free kill" based on some sort of avoidable ability, then you will have hard times.


Also I'm pretty sure "wins" taken into account. Friend of mine started DotA few months ago. He has 100 wins or so and all other players in games have similar wins.

[Sackman]

Can you explain how did you manage to solo kill someone in lvl5 in middle(safest lane) with shadow demon? Because I am having hard times to believe your opponent was a good one.

Also do you got some sort of argument-o-meter pluged-in to you? Because every one in here is basically doing same thing; saying this hero is better on bla bla and providing arguments. You seem to be only one who complaining about the quality of arguments. I am guessing you are measuring it. Not that I don't find every argument okay, but when I don't, I provide more counter arguments instead of complaining about it every two pages.

Anyway, there are way better alternatives when it comes to ganking, burst damage or lane control compared to SD. He's an excellent support but not a solo laner imo.

I've played solo mid and got kills with SD, but only when opponents make mistakes. Fortunately, people make mistakes all the time, it's human nature, and thus SD is a viable solomid if you presume your enemy is not infallible.

Alongside that, he is a great roamer, let's say, UD hardlane, Jugg+Veno easy, NP jungle, SD mid. SD is an easy mid, with great range, and great gank potential. Alongside that, disrupt works as a viable escape, nuke and disable, and his ult means even heros like windrunner are not immune.

With that being said, there are many other choices, like using the same heros you could put veno mid instead and it would work similarly. Its all up to you whether you want to change up the way you play. I think thats the problem with the pro scene atm, nobody changes anything up, despite there being multiple viable builds.

[Hermanni]

BAM! Statistical theory proved you wrong!

I read your post 3 times yet I can't find the part with statistical theory proving anyone wrong. Or anything else that even remotely looks like a proof. Or any kind of statistical theory even making appearance. Or any sign of him being wrong in the first place.

Bam, logic, boom, your post lacks some.

[Sackman]

Yep, same for me. I always laugh when I look up people I know on dotabuff and look at their most played with list. They have a bunch of friends that they have 80% wins with and I look at mine and every single friend I play with is negative. Must be nice to win all the time when you queue with friends =p

I think the MMR responds to your level tbh. Most games I play now involve people with 600-1000+ wins, which is a world away from where I was when i started (20-40 wins).

That being said, I'm at about 270 wins, which would lead me to believe my MMR must be pretty high if I am constantly playing with people who have up to 5 times as much experience as me. (I never played Dota1 and this is basically my first moba, although I did play LoL to level ~14 before I got an invite, LoL is so easy all it did was teach me some nomenclature).

[Kuntantee]

I read your post 3 times yet I can't find the part with statistical theory proving anyone wrong. Or anything else that even remotely looks like a proof. Or any kind of statistical theory even making appearance. Or any sign of him being wrong in the first place.

Bam, logic, boom, your post lacks some.

It's easy.

Games are played by players. Statistical data and results collected ignoring the human factor because human skill also decide the winning side. You can have a truely overpowered hero but still lose. why? Because enemy team somehow outskilled you be it on picks or some other way. What I am trying to say there is no statistical way to find whether a game is lost because enemy team had an OP hero or they were plain better skilled. So all those statistics, again, ignoring human factor which can not be ignored. Also statistical data collection suggest that there should be only one variable(irrelevant variables can be ignored) and that's what you are trying to find.

In this case, teams are different(which is a vast difference) and other heroes are different(another vast difference). Statistics is not applicable here. There are special methods that involve 2 at max 3 variable(not random variables tho) but you have tons of variables here.

Before you trying to mock me with logic etc. First look at the mirror and see someone who fails to understand high school mathematics.

[Deleted]

It's easy.

Games are played by players. Statistical data and results collected ignoring the human factor because human skill also decide the winning side. You can have a truely overpowered hero but still lose. why? Because enemy team somehow outskilled you be it on picks or some other way. What I am trying to say there is no statistical way to find whether a game is lost because enemy team had an OP hero or they were plain better skilled. So all those statistics, again, ignoring human factor which can not be ignored. Also statistical data collection suggest that there should be only one variable(irrelevant variables can be ignored) and that's what you are trying to find.

In this case, teams are different(which is a vast difference) and other heroes are different(another vast difference). Statistics is not applicable here. There are special methods that involve 2 at max 3 variable(not random variables tho) but you have tons of variables here.

Before you trying to mock me with logic etc. First look at the mirror and see someone who fails to understand high school mathematics.

This just in: Forum poster disproves entire field of mathematics, what does this mean for the world?

[Sackman]

It's easy.

Games are played by players. Statistical data and results collected ignoring the human factor because human skill also decide the winning side. You can have a truely overpowered hero but still lose. why? Because enemy team somehow outskilled you be it on picks or some other way. What I am trying to say there is no statistical way to find whether a game is lost because enemy team had an OP hero or they were plain better skilled. So all those statistics, again, ignoring human factor which can not be ignored. Also statistical data collection suggest that there should be only one variable(irrelevant variables can be ignored) and that's what you are trying to find.

In this case, teams are different(which is a vast difference) and other heroes are different(another vast difference). Statistics is not applicable here. There are special methods that involve 2 at max 3 variable(not random variables tho) but you have tons of variables here.

Before you trying to mock me with logic etc. First look at the mirror and see someone who fails to understand high school mathematics.

Statistics and figures practically NEVER actually provide a truth, they are there to support a hypothesis. Which is why people refute your argument, you have an argument, but provide no evidence, which just makes it an opinion.

For example, in my experience, Drow is op (my opinion), if I check https://dotabuff.com/heroes/winning, I can see she has the highest win rate and has quite a comparable distance on other hero's (who are arguably also op, looking at you jakiro/centaur) this month, which supports my claim.

I believe this patch made her op, so what can I do to support this statement? Compare patches:

6.76c: 59.35% (the minor nerf to her ult)
6.76: 60.58%
6.75: 51.21% (the patch before).

Personally, I don't think the buff she received was inline with the rest of the game, and she has become unduly powerful. I have supported this with statistical data. This makes my argument "drow is op" much more of an argument rather than an opinion.

[Hermanni]

Before you trying to mock me with logic etc. First look at the mirror and see someone who fails to understand high school mathematics.

Can I borrow your mirror?

No, seriously, I haven't seen you present anything that even remotely looks like mathematics, just random rambling and some lingo. I can assure you I'm very learned in maths, high school level and beyond, and statistics and probabilities were always one of my favorite fields.

[Lysah]

No one would disagree with Drow being OP right now, you don't even need win rate %s to back that up. 200 damage at level 6 naked herp

[NotValidAnymore]

http://pastebin.com/Q4E3HDCE

It is time.

Sorry to break off the conversation in here, but if I'm not reading wrong, is that the guy from Bastion as an announcer in Dota 2? I think my mind just jizzed.

[Finear]

gief GLaDOS announcer pack <3

[Kuntantee]

Can I borrow your mirror?

No, seriously, I haven't seen you present anything that even remotely looks like mathematics, just random rambling and some lingo. I can assure you I'm very learned in maths, high school level and beyond, and statistics and probabilities were always one of my favorite fields.

You are acting the smart guy in this thread, criticizing the quality of arguments yet you can't do any better(love the part you quoted my post and failing to provide a good counter argument other than saying its rambling and lingo ). You are failing to understand what my post is trying to say. You don't need to know mathematics to understand winrates provide no useful information. It's very basic logic as it can get. I am fairly sure my little cousin has better understanding than you do.

I could explain why winrates are not useful but that would be my 10th times repeating myself You can keep circle jerking while thinking how awesome arguments you guys provided with winrates. I am done.

---

I hope Valve won't force us to hear other's announcer pack...HoN's system is terrible.


If you can't post without insulting another user, don't post at all.

[Finear]

I hope Valve won't force us to hear other's announcer pack...HoN's system is terrible.

you can choose which one you want, i see no reason why would they change it