[Megathread] DotA 2

[Longview]

Edit: According to Worer's link, Morphling had a 61% winrate as a hard carry, and 81% winrate when he was solo mid.

You see how easy it is to back up your claims? Would you argue that Morphling was too strong in the middle lane? As a hard carry? You now have something to back this up rather than "1v4 GG" or "too strong". This gives your post credibility, and people will be more likely to think of them as actual arguments from someone who knows what he is talking about, rather than inane babble.

I'm not skipping posts, I'm just not quoting them or responding to them. I suppose that is skipping depending on how you look at it. Anything in specific you would like me personally to respond to? Out of curiosity, what evidence did you provide that shows how stats in public games do not matter? I'm not refuting your point, just interested.

[Deleted]

You see how easy it is to back up your claims? Would you argue that Morphling was too strong in the middle lane? As a hard carry? You now have something to back this up rather than "1v4 GG" or "too strong". This gives your post credibility, and people will be more likely to think of them as actual arguments from someone who knows what he is talking about, rather than inane babble.

I'm not skipping posts, I'm just not quoting them or responding to them. I suppose that is skipping depending on how you look at it. Anything in specific you would like me personally to respond to? Out of curiosity, what evidence did you provide that shows how stats in public games do not matter? I'm not refuting your point, just interested.

I wrote the 1v4 as an example of how maybe someone could think it is fine since both teams can pick him up, then it's "balanced" and he can be banned.. I have heard that argument before.

Hermanni's argument was that Morph was fine before since he had 53% winrate in pubs, and now to horrible 38% winrate in pubs. Then I said that AM and Broodmother who are both competitive are actually only right above him with a slightly higher winrate. And just to come with another example, Magnus who won all but two games in DHW has a 48% winrate in pubs.

Then Hermanni said that only Morph was so bad that he is terrible in pubs, and doesn't get to see competitive play(Meaning that either a hero is good in pubs, in compeitive play or both). Then I gave a list of other heros that probably see less competitive play than Morph(who I saw picked in G1 as a serious pick, by Orange I believe?), while still having slightly-to-far less than average 50% winrate in pubs.


And I don't think the morphling stats are legit, since EVERYONE played him as a hard carry, and 80% of the games was solo mid. Then why the low 52% overall winrate, but when on those positions(hard carry, solo mid) he has insane high winrates? Only stats of one game playing as position 2, which he won..There are tons of games missing, unless they only record his role and lane position in some games? Then which? High-end compeitive ones like TI2?

I think the stats are off, that's why I didn't mention it.

To close this off I am actually not saying Morph didn't get nerfed too hard, but I am sticking with he needed a nerf.

[Kuntantee]

Win rates, both in pubs and professional scene, mean nothing. I don't see how that proves anything. Actually that would mean a lot if game run by some sort of AI but, well, it's played by different players with different skills and there is also something called "strategy".

[Lysah]

Win rates by specific heroes only matter if you start noticing a trend in first pick win percentage going up. If first pick wins ~51% of the time and loses ~49% of the time, individual heroes probably don't matter so much. If a hero like Magnus ends up having a 95% win rate when picked and you suddenly notice that first pick wins 60-65% of the time, and, oh hey, Magnus was picked first a lot, you have a problem.

[Longview]

Win rates, both in pubs and professional scene, mean nothing. I don't see how that proves anything. Actually that would mean a lot if game run by some sort of AI but, well, it's played by different players with different skills and there is also something called "strategy".

They average out, so they do matter. The winrates of teams are more interesting than the winrates of heroes, though. The Radiant has a higher win-percentage in professional games. This is largely attributed to the Radiant most often having first pick. Since 6.75, the Radiant has achieved 162 wins compared to the Dire's 129 (166/126 (first/second pick respectively)). Hero winrates matter little when you do not divide them into skill brackets, however. I suppose you do have a point because win rates alone say nothing important.

[Lysah]

Hero winrates matter little when you do not divide them into skill brackets, however. I suppose you do have a point because win rates alone say nothing important.

What do they mean even if you divide them into skill brackets? We can say "Ursa is balanced because he doesn't do well in pro games and is never picked," but I think most people would agree that Ursa stomps average skill public games which are 90% of the world. Is it right that we balance around the professional scene and the vast majority of everyone else can just deal with it? Or do we balance around the majority at the expense of the esport scene? Ideally, we'd like to find heroes who perform well in all brackets, looking at win rates going up or down as the skill division increases generally just gives you a graph of how difficult a hero is to play.

[Longview]

They do not mean much individually (I would however really enjoy to see such statistics), but I don't fully agree with the last part of your post. From my (somewhat limited) experience, most heroes are pretty straight forward to play. Sure, they're not all as intuitive as Skeleton King or Ursa, but I wouldn't say there's that many heroes which could not be played to at least a passable degree by your average public player. I guess what I'm trying to say is most of the heroes are fairly easy. Across the board win rates get incredibly skewered when you do not see your average pub stomping heroes in professional games (and vice versa), hence them not being a good indicator of how strong a hero is.

I had this discussion earlier with the captain of my clan, it's interesting. I personally don't think that the game should be entirely balanced around the professional scene, or even the high-tier players at large (hence why captains mode is great imo). I can't speak much for public games, as most games I play are lobbies - and the games I watch are either replays or progames/public games with progamers. It does make sense to balance the game around the most average rating, but also around the highest. I am just happy IceFraud is doing his thing..

[Lysah]

So far so good I suppose. Now if he can just get the queue times under fifteen minutes for me...

Trying to keep the game competitive for the spectator sport part of the game and still easy to learn and fair for the entry level players is a hard job, you see this struggle with Riot trying to pump heroes out of the machine every few weeks. How do you keep a hero "shallow" enough that anyone can pick it up and learn it in a few days without making it so easy/good that it changes the entire game? I don't envy that part of design.

[Deleted]

So far so good I suppose. Now if he can just get the queue times under fifteen minutes for me...

Damn.. I only queue for europe and I never get over 5min queue. Earlier today I got it instant even. O_o

[Lysah]

Damn.. I only queue for europe and I never get over 5min queue. Earlier today I got it instant even. O_o

When I solo queue I usually get 3-5 min, that's okay I guess. When I queue with friends the system gets confused or something because we never get a game until the search range is maxed at 10 minutes.

[PizzaSHARK]

So far so good I suppose. Now if he can just get the queue times under fifteen minutes for me...

Trying to keep the game competitive for the spectator sport part of the game and still easy to learn and fair for the entry level players is a hard job, you see this struggle with Riot trying to pump heroes out of the machine every few weeks. How do you keep a hero "shallow" enough that anyone can pick it up and learn it in a few days without making it so easy/good that it changes the entire game? I don't envy that part of design.

Riot pumps out new heroes because it's the basis of their income scheme. Valve doesn't have to do that because you can buy hats and couriers and tournament tickets.

The difficulty is there because they want to push Dota 2 as a competitive eSport as well as make it a cash cow for more casual players. The issue is that Dota requires an enormous amount of information and experience for you to do even semi-well at, which makes it hostile to newcomers almost by default.

Tutorials will be able to handle this, but honestly... there's an extent to which tutorials can do anything. I was teaching a buddy about Dota and we spent six hours straight playing botmatches and focusing on different things in each match and I still didn't think he was anywhere near ready to play against other players. Maybe playing with other players against bots, but not against other players.

That's a frankly ridiculous kind of time investment to not even be at the bottom level of performance, and since pub Dota especially is almost purely based on who has the most games played... it's gonna be a very tough challenge to make the game truly interesting to newbies. I expect most people will try it since it'll be F2P and abandon it when it becomes obvious they can't spend 30 minutes in tutorials and come out knowing how to play.

[Kuntantee]

win rates alone say nothing important.

That's exactly what I was trying to say. Win rates provide little for analysis if you do not consider other relevant information. Also before first pick there are 3 bans and dire has two picks straight. It just increases the strategy cap imo. First pick of radiant comes into mind but I really doubt that is the reason behind such difference. Actually it may be the case but proving it would require so much computation(that is only applicable if heroes are truely balanced and both team, somehow, have equal skill). I'll just call it luck.


Why do I talk about proving? Since we are talking about numbers and statistics proving some sort of claim would require the reliability of provided information. Especially if its in numeric form. Since there is no way to prove relevancy of such claims with some poor statistics(unless you prove it mathematically which can not be the case). I just do not consider any of these websites as a usable source of information for in-game analysis.

Balancing the game for casuals would ruin my experience imo. This is a personal choice. I just take DotA 2 seriously and try to improve myself. In the end, I have enough experience to counter current and future pub stomping heroes like Ursa or Lycan and I can't imagine the state of Lycan or Ursa balanced for pubs. They would be somehow underpowered. I do think that this game may eventually be shaped by non-competitive scene because if you can't(not bad players) keep casuals happy, you just can't get enough profit. This may not be the case as well because DotA has a strong non-casual(at the same time non pro) player base.

Time will tell.

[Lysah]

@pizza
I agree, will be interesting to see how it plays out. LoL is definitely friendlier to newbies.

Also, as far as radiant vs dire win rate goes, the map is not a mirror and I believe the radiant side is slightly stronger than dire.

[PizzaSHARK]

@pizza
I agree, will be interesting to see how it plays out. LoL is definitely friendlier to newbies.

Also, as far as radiant vs dire win rate goes, the map is not a mirror and I believe the radiant side is slightly stronger than dire.

I've been told Dire's jungle is better, and they definitely get easier access to Rosh. What major advantages does Radiant get that can more than make up for those kinds of advantages?

[Finear]

What major advantages does Radiant get that can more than make up for those kinds of advantages?

middle lane pulling

but i was under impression that dire side of the map is better
afaik dire always had higher w/l ratio (although by very small margin)

[NotValidAnymore]

Ugh... what is with all the new people and russians I get on my team in PvP? I mean, I can see why, if I've won a lot of matches, I would get some newbies on my team, but in that case the other team should as well.

So far I've already had a couple of matches where I had feeding mids who never gank, "pre-made" two-man teams who don't call miss and the other team roaming and killing everything. Their side? Oh, they know what they're doing, every time.

I guess, like LoL, that whenever I play a PvP match, I'll have to take it with a grain of salt and get used to 3-4 random carries who are basically useless.

[Deleted]

I've been told Dire's jungle is better, and they definitely get easier access to Rosh. What major advantages does Radiant get that can more than make up for those kinds of advantages?

Dire has:

Better pulling from easy lane. (Actually three potential places)
Huge rosh advantage.
Easier ancient stacking stacking for off-laner.
Ancient pulling to lane with certain heros(Venomancer atleast).

Radiant has:

Pulling from mid lane.
The one pull they get in easy lane is more effective, but overall worse than what dire can do. You can pull the creep wave into three camps and get more xp, while dire takes out two camps max. per pull, but can clear 4 camps total with 2 pulls
Two safe jungle camps. What I mean is the big and small camps towards middle lane are pretty much always safe to farm until both tier 2 mid/bot goes down.
Ancient stacking from middle laner, while still being viable to stack from off-lane, but a lot more of hassle than for the dire side.
Off-laner can creep pull into dire jungle.

Probably more stuff, but this off the top of my head :P

[Longview]

That's exactly what I was trying to say. Win rates provide little for analysis if you do not consider other relevant information. Also before first pick there are 3 bans and dire has two picks straight. It just increases the strategy cap imo. First pick of radiant comes into mind but I really doubt that is the reason behind such difference. Actually it may be the case but proving it would require so much computation(that is only applicable if heroes are truely balanced and both team, somehow, have equal skill). I'll just call it luck.

Since 6.75, there are two bans in the first phase of drafting for each side. Not three.

Why do I talk about proving? Since we are talking about numbers and statistics proving some sort of claim would require the reliability of provided information. Especially if its in numeric form. Since there is no way to prove relevancy of such claims with some poor statistics(unless you prove it mathematically which can not be the case). I just do not consider any of these websites as a usable source of information for in-game analysis.

If you perform a simple chi square test, you will find that there is a significant difference since 6.75 (p<0.01). Before the patch, win rates were lined right up at 50/50. You could always argue that there are too many random variables, but you can take any 30 day period prior to 6.75, and find the win rates be 50/50. Suddenly there's one side winning 10% more games than what the other side is. But you just call this luck, right?

If we, for the sake of the argument, assume that this is not random - the question becomes: why? I do not see any other reason than first pick being that much more potent. The team that has second pick always get their strong two, and the seldom used surprise last pick. I personally prefer drafting as first pick, it feels like our draft end up better than theirs more often than not. But that's just what I feel. Are there any other reasons bar luck and firstpick being that good for the sudden burst of Radiant wins? Aegis six minute fade?

[Lysah]

Radiant has a safer jungle, Dire is easier to invade because of the location of their wards and the vision they get from them. Radiant can cover every entrance with one ward and Dire cannot. Dire has to cross the river to place offensive wards to protect their jungle and Radiant does not. Both teams have about equal jungle pulling, Radiant can pull mid and can also triple pull bot VERY easily, double/triple pulling dire is a lot harder and uses the small camp which is not as useful exp/gold wise.

Biggest difference for me is ancients. Radiant can protect their ancient camp much more easily than Dire due to proximity to mid. You can call it balance that Dire's ancients are harder to invade due to vision, however, radiant doesn't have to fight uphill the same way Dire would if they tried to stop a stack. Dire certainly gets an easier roshan - does this make up for the weakness of their jungle, I'm not sure. These differences aren't game breaking, but they exist. When I played DotA1 I recall Sentinel having a slightly higher winrate pretty much always, Dire being stronger in dota2 is news to me. I would be willing to bet Roshan is the difference here, as many professional players win and lose games around this now when he wasn't really a big factor in dota1.

[Longview]

On the contrary, jungling inside of the Dire jungle is more efficient, and their top T2 tower is much more difficult to bring down than the Radiant counterpart. You'd think the win-rates would even out as people adapted to 6.75, but if you filter more recent games - they're even more skewed.