|
United States47024 Posts
On June 14 2014 17:14 phantasmal wrote:The interesting thing KF91 has pointed out with the Summit has nothing to do with the side stats, which never would have been large enough to be relevant on their own, it's how the finals drafts appear to have been influenced by the idea of the Dire late-game advantage. It's not even the first time that DK has gone full push on the Radiant in a major finals this year, but I had a lot more confidence in that lineup than the one they used in game 2 vs EG.
The thing is, if you're going to discuss how being Radiant pressures a team into drafting a certain way, you also have to have some mention of how first pick/2nd pick does the same. Being 2nd pick Radiant is sort of a compounded disadvantage because of how being first pick ALSO gives the Dire the initiative with regard to the draft, which could be what forces a Radiant 2nd pick into a situation where they feel they must take a certain draft.
|
I pulled the International 2013 statistics, which may well be out of date by now, and crunched some numbers.
158 games, 82 wins by radiant, 76 wins by dire. The difference between average win times was about a hundred seconds (Dire 36:33, Radiant 34:48, all games 36:38 ), with standard deviations over ten minutes either way (Dire 11:56, Radiant 13:26, All games 12:43). It's safe to assume that there wasn't much difference back then.I think it's safe to assume that there wasn't a difference back then.
Of course, the same for this tournament gives a difference of 12:55 in average win times (Dire 44:30, Radiant 31:34, all 39:15) with standard deviations about the same (Dire 11:41, Radiant 10:20, all 12:45). Welch's t-test, unless I messed up somewhere, gives a value of about 3.5, which means around 0.1% odds that there isn't a difference.
|
I think changing the bans in the last two games by DK was a throw. if they had just first banned tree and enigma like in the first few games they would have crushed them to win the summit. I Believe DK throwed 1st place because of the draft.
|
276 Posts
On June 15 2014 02:34 TheYango wrote:
The thing is, if you're going to discuss how being Radiant pressures a team into drafting a certain way, you also have to have some mention of how first pick/2nd pick does the same.
No you do not, because..
Being 2nd pick Radiant is sort of a compounded disadvantage because of how being first pick ALSO gives the Dire the initiative with regard to the draft, which could be what forces a Radiant 2nd pick into a situation where they feel they must take a certain draft.
A) Calling 2nd pick Radiant a "compounded disadvantage" assumes that Radiant is a disadvantage, which has not been proven. Radiant has had the net advantage in multiple patch periods and in pub play every time it has been looked at. I still wouldn't call this proof that Radiant has a net advantage, but it certainly ought to disprove that there's an automatic Dire advantage.
B) 2nd pick might influence drafting in some way, but until evidence is provided it should not be assumed that it does influence drafting. 2nd pick disadvantage in certain periods could be as simple as "Game has 5 really strong heroes, and 1st pick is significantly more likely to get at least one of those 5," in which case there would not really be a lot to adapt to as a drafter besides working to make sure your first two bans are good ones.
C) The Radiant/Dire time balance will influence your game regardless of whether you're first or second pick, and if teams are taking it into account during their drafts as 2nd pick on Radiant there is no reason to believe that they would suddenly stop taking it to account as 1st pick. For instance, the game I linked earlier was a 1st Pick Radiant for DK where they went Pugna push. At best you might eventually make the point that some subset of drafters are more likely to go for early game or push strats as 2nd pick Radiant than they are as 1st pick Radiant, but that would be a complicated and massive statistical study that goes well beyond the scope of this article.
If this were a side balance article, then yes, first pick should have been mentioned. And thankfully it's not, as I would have killed riptide for making me read through something as banal as a side balance article for a 37 game tournament. KF pointed out a behavioral response to an effect that exists regardless of pick order. And you know, maybe there is also a story about pick order somewhere out there, but that's a separate article for a separate time.
|
it seems to me to be clear that both DK and EG are just better at playing on dire, perhaps they're better at taking rosh than preventing it, better at defending a ramp than taking it, and so on.
|
Well, one side might have the advantage, but, do they still keep doing the "pick sides" thing? Like, the winner of a flip coin or w/e chooses side, and the other chooses 1-2 pick?
If that's the case, then it depends on the players, more than this. If you can select the side, then don't blame the sides if you lose.
|
On June 15 2014 11:08 linkhimura wrote:
Well, one side might have the advantage, but, do they still keep doing the "pick sides" thing? Like, the winner of a flip coin or w/e chooses side, and the other chooses 1-2 pick?
I believe zai or ppd mentioned there was a coinflip to determine who would get to pick the side in the 5th game. The first four games were alternating between Dire and Radiant. I'm not sure how the first pick worked since Dire ended up having first pick 4 out of the 5 games... Maybe someone who actually remembered how they set the rules chime in? 
If that's the case, then it depends on the players, more than this. If you can select the side, then don't blame the sides if you lose.
I don't think that they were blaming sides, they just had different strategies depending on which side they were on.
|
Some good meat on this article 
It's cool to see radiant and dire being different. Gives extra depth to the picking stage. I wouldn't call dire to have an advantage, winning early is a valid strategy and radiant is good for that.
Hertzy, for the small sample of this tournament, you consider only dire vs radiant and match times, but don't take picks into account. New Meta might show up in TI4. On that note, would be cool to see a followup on this story then.
|
On June 14 2014 05:03 Sn0_Man wrote:
I didn't see you mention it, so I'll say this: the reason radiant do so well in short games is because their offlane is much much easier.
Just thinking about radiant offlane gave me a boner - it's THAT good!
In all seriousness I'd also consider the ancients/neutral camps when talking radiant vs dire. Having played a lot of Tinker and jungle Bat over the last three months, I can definitely say that radiant is much better to farm and both also seems much more secure.
|
On June 15 2014 19:34 Badjas wrote:
Hertzy, for the small sample of this tournament, you consider only dire vs radiant and match times, but don't take picks into account. New Meta might show up in TI4. On that note, would be cool to see a followup on this story then.
Yeah, that wasn't in the .csv datDota provided, for some reason. It takes two clicks per match to find that out, so it'd take a while to compile that data.
EDIT: I probably missed a match or two, but I see Radiant being 2/9 when Dire goes first in draft, and 11/27 when Radiant goes first in draft.
EDIT2: Unless I screwed up, chi-squared test for independence gives p-value 0.31, which means you'd expect to get this result in one in three tournaments even if going first in draft didn't affect the outcome.
|
What impact does first pick have on the outcome of the game as an isolated variable and in conjunction with map side (radiant/dire)?
|
Radiant has 40% winrate while going first, 22% when going second, 1.9-fold difference. Dire has 78% winrate when going first and 59% when going second, 1.3-fold difference.
This is with p-value 0.31, however, and traditionally it would have to be less than 0.05 for this to be statistically significant.
|
I just looked at the scrap of paper I made the count on and noticed that, minus the one match I missed, first picker had exactly 50% win rate.
|
Doesn't Radiant currently hold an overall greater win% in both pubs and competitive play, since the game average time is 35 minutes roughly and the Radiant has a higher winrate in that timeframe?
|
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced.
|
On June 17 2014 08:11 mutantmagnet wrote:
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced.
I don't think this is related. The current speculation AFAIK is that radiant has safer ancients with faster access, a safer jungle with faster access, and their off lane can lose a tower without too many issues while the Dire losing their offlane tower eliminates a lot of map control. Dire gets Roshan in exchange for this, and in the very late game it seems to favor them for that reason.
|
On June 17 2014 08:11 mutantmagnet wrote:
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced.
Like you said, it's not the same game so I'm not sure if that reasoning would hold over different games... But this is pretty interesting. I know this is a Dota 2 forum, but would you mind sharing a link? It would be interesting to see why Radiant had a ended up having a higher win rate. I tried searching for "League of legends symmetry/symmetry mode/symmetric map" but nothing came up..
|
On June 17 2014 08:23 KF91 wrote:Show nested quote +On June 17 2014 08:11 mutantmagnet wrote:
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced. Like you said, it's not the same game so I'm not sure if that reasoning would hold over different games... But this is pretty interesting. I know this is a Dota 2 forum, but would you mind sharing a link? It would be interesting to see why Radiant had a ended up having a higher win rate. I tried searching for "League of legends symmetry/symmetry mode/symmetric map" but nothing came up..
The camera advantage isn't proven to exist AFAIK, at least not in Dota where the camera is freemoving. I suspect the advantage in Dota is purely down to map control and map safety, which favor the Radiant for a match of average length, until the late game where Roshan positioning and even map control begins to favor the Dire to take the Aegis.
|
On June 17 2014 08:27 Zoolooman wrote:Show nested quote +On June 17 2014 08:23 KF91 wrote:On June 17 2014 08:11 mutantmagnet wrote:
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced. Like you said, it's not the same game so I'm not sure if that reasoning would hold over different games... But this is pretty interesting. I know this is a Dota 2 forum, but would you mind sharing a link? It would be interesting to see why Radiant had a ended up having a higher win rate. I tried searching for "League of legends symmetry/symmetry mode/symmetric map" but nothing came up.. The camera advantage isn't proven to exist AFAIK, at least not in Dota where the camera is freemoving. I suspect the advantage in Dota is purely down to map control and map safety, which favor the Radiant for a match of average length, until the late game where Roshan positioning and even map control begins to favor the Dire to take the Aegis.
That suspicion (along some of my supporting ideas in my text) was defied today with the closing of the [A] vs C9 finals where Radiant not only won all 5 games, but the average game length was much longer (42 minutes)...
I haven't taken a deep look into the drafts/builds yet, so maybe there was something different that I didn't see (although EE had some very... questionable plays today)
|
On June 17 2014 08:49 KF91 wrote:Show nested quote +On June 17 2014 08:27 Zoolooman wrote:On June 17 2014 08:23 KF91 wrote:On June 17 2014 08:11 mutantmagnet wrote:
Radiant has the clear advantage.
It isn't the same game but the results are irrefutable.
Recently League of Legends came out with a map mode where everything was symmetrical and over time radiant has been winning 60% of their games.
Dire side needs to be imbalanced. Like you said, it's not the same game so I'm not sure if that reasoning would hold over different games... But this is pretty interesting. I know this is a Dota 2 forum, but would you mind sharing a link? It would be interesting to see why Radiant had a ended up having a higher win rate. I tried searching for "League of legends symmetry/symmetry mode/symmetric map" but nothing came up.. The camera advantage isn't proven to exist AFAIK, at least not in Dota where the camera is freemoving. I suspect the advantage in Dota is purely down to map control and map safety, which favor the Radiant for a match of average length, until the late game where Roshan positioning and even map control begins to favor the Dire to take the Aegis. That suspicion (along some of my supporting ideas in my text) was defied today with the closing of the [A] vs C9 finals where Radiant not only won all 5 games, but the average game length was much longer (42 minutes)... I haven't taken a deep look into the drafts/builds yet, so maybe there was something different that I didn't see (although EE had some very... questionable plays today)
It's an average effect, not a prediction. You can't look at 5 games and declare the outcome means anything about the whole system.
|
|
|
|
|
|
EPT
