The Radiant Advantage

Interesting post here about a 59%-41% advantage for Radiant in league matches since 6.75: Did 6.75 disrupt the balance between Radiant and Dire?

The two big relevant changes in 6.75 as described in the post were the shortened Aegis reclaim time and the realignment of the ban and pick phase.  My first thought is just do a test of 6.75/76 matchmaking games.  If the trend isn’t repeated then the most likely culprit is the pick/ban change.

Unfortunately I can’t do this while the API is still down, but it occurred to me that I had never done a Radiant vs Dire test on my own 6.74 sample.  So I threw a remarkably quick one together and came across some interesting results.

First, we have the Radiant win rate for all games in the sample (with a > 15 minutes duration check to purge out most of the games ruined by abandons)

Norm/High/VHigh

51.6/53.3/54.0

Sample Size:

9833/9915/9829

An interesting shift.  If you account for the fact that Normal games are 4 times more common than High+Very High, the adjusted average Radiant win rate is very close to 52%, but the advantage seen in the top two brackets is considerably higher.

So what happens if we only look at games that last at least 30 minutes

Norm/High/VHigh

51.1/52.0/52.3

Sample Size:

9024/8433/7874

Much of that Radiant advantage dissipates in longer games.  Does this trend continue indefinitely?  Let’s look at games that last 45 minutes or longer:

Norm/High/VHigh

48.3/49.4/49.7

Sample Size:

4496/2967/2473

If anything, the trend away from the Radiant advantage appears to accelerate as the games go longer, so let’s go the opposite direction and look at games that are shorter than 30 minutes.

Norm/High/VHigh

56.9/61.0/61.1

Sample Size:

787/1457/1930

The trend continues in dramatic fashion.

So what can we take from all this?

  1. For starters, whatever advantage Radiant has is strongest in the laning phase.  The fact that this erodes away as the games draw longer suggests that Dire has a late game advantage, which is consistent with the theory that the location of the Rosh pit favors Dire.
  2. But unless the game goes longer than approximately 45 minutes, Radiant’s advantage is stronger than whatever benefit the Dire get from Roshan.
  3. Both the Radiant early advantage and the Dire late advantage appear more significant in the High and Very High brackets.
  4. Valve may or may not have a built-in matchmaking advantage for either side.  For example, there could be a pro-Radiant matchmaking handicap designed to compensate for the Roshan advantage.  If this handicap overwhelms the Dire during laning then Rosh doesn’t matter.  If it doesn’t overwhelm the Dire then the game becomes a toss-up.

No definitive conclusions can be reached other than to say that match duration appears to play a pretty significant role in Radiant vs Dire balance.  Hopefully future samples and tests will be able to expand on this base.

Advertisements

5 Responses to The Radiant Advantage

  1. xdv says:

    That Reddit link suggests that that mid pulling is responsible for the Radiant lane phase advantage but I don’t recall seeing that happen a lot in the last international…. when you say this sample size is from “league” games you mean tournaments like those at the TI2, yeah?

    If anything I feel Dire even has the much stronger laning phase where you can pull the safelane easy, medium, and hard camp (and the hard camp is even autostacking during the pull) while a single ward can deny the radiant safelane any pulling options, but that’s just one of the many asymmetries in the map…

    • phantasmal says:

      The Reddit sample is, if I understand them correctly, every game recorded by dota-academy since 6.75. So yeah, it should be every major tournament in the last ~2 months, give or take a couple days.

      My sample is public matchmaking games in 6.74. I don’t know the distribution of the game modes, but I suspect that All Pick is the majority of the sample.

      They do suggest that mid pulling might play a roll. If this is true, it’s probably a recent phenomena. I doubt this makes a very large difference in matchmaking, and am almost certain it makes no significant difference in normal matches.

      Basically there are 3 (at least) possibilities as to why one side would have an advantage:

      1. Pick order
      2. Map differences
      3. Matchmaking rating adjustment

      For tournament games 1 and 2 would be relevant. For matchmaking games 2 and 3 would be relevant.

      So for me, there are two big missing pieces to this puzzle.

      1. Is there a Radiant advantage in public matchmaking games in 6.75/6.76c? Knowing this let’s us compare pre-patch and post-patch matchmaking games AND ALSO compare matchmaking games to tournament games within the same patch. The latter could be particularly important for estimating how much of a roll pick order plays in this.
      2. Duration appears to play a very important role. Does this trend hold true in tournament games? Dota-Academy has 1833 games in pre-6.75. If we find the median match duration, does Radiant have a higher win rate in games that are shorter than the median duration? Either answer could have some interesting implications.

  2. Steve says:

    I wish there was some way to compare the laning phase, but with the data you have available I don’t see a good way. I’m not sure if this is possible, but one thing I thought of was to check only for games that contain both a common jungler and common solo off-laner. Maybe then with enough of these games you could compare how the jungler and off-laner each perform on average for each side of the map. Maybe it could have some significance with enough data.

    • phantasmal says:

      I’m inclined to share your fear that there’s no good way to get at that kind of information with my data, but what might work is something using the demo parser from http://dev.dota2.com/showthread.php?t=32672

      The demo files should contain a bunch of information like hero deaths and creep kills, plus the game time and location for each event. Define area boxes for each lane, and then two big remainder boxes for the jungles, and then measure the average creep kills and deaths for each side in the first 15 or 20 minutes to find discrepancies. If it’s really good it might even be able to detect and sort by team lineups, like 2-1-2, offensive or defensive trilanes, etc.

      But it’ll have to come from someone else. So far I’m in way over my head when I mess with the demo parser.

      Two other cool things the demo parser should allow that aren’t available from the API are measuring rune and Roshan control.

  3. shostakovich says:

    Hey! Nice read, thank you. Any chance we can talk? Send me an e-mail with skype contact or something like it, brunolopestomaz@hotmail.com

    Cheers!

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: