Here is a quick and dirty look at twoeye's record, comparing: * The 100-game moving average winning rate = "actual" * A smoothed version of the same data for easy reading = "forecast" (excel's built in "exponential smoothing" feature) * The dan level multiplied by 10 = "rank" (thus 6d = 60 in the graph; just to put it easily in the same graph) We can clearly see that changes in rank sharply affected the winning rate.
twoeye winning percent vs rank.jpg (160.49 KiB) Viewed 3526 times
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21
ETA: Also ez4u, you just took win/loss data from sum's history? Going how far back? His rank has changed from 5d to 4d 3 months ago. Throwing games played as 5d and games played as 4d together will confuse things a lot I think.
I think this is a good point, but I don't know how good.
The table below shows the winning record of the three bigs broken down by their rank at the time the game was played (our speedster lies on a completely different scale and so does not make it into this graph). Each of the three has two different ranks making up significant portions of their records, with clearly different winning percentages. This alone should force the issue of non-stationarity into the statistics if I understand that concept correctly.
BTW, I downloaded and used the complete kgs history of each player. They appeared on kgs (in their current username anyway): * thecaptain 2002-09-26 * sum 2004-03-25 * twoeye 2004-05-12
Wins by Rank Summary.jpg
Thinking more about yoyoma's point, I built a little simulation in excel based on twoeye's record. I set up 11,000 cells to randomly generate wins and losses at a 66% win rate together with 4,000 cells to randomly generate wins and losses at a 54% win rate and passed the results through the Wald-Wolfowitz runs test. I wanted to see whether this would reproduce the kind of "streakiness" (lower than expected number of runs) that we can see in the actual data. Running it 100 times gave an average (mean) Z statistic of -1.64, which is less than we saw with our big three but is out at around the 95% confidence level for rejecting the hypothesis that the result is random! In the 100 runs, the maximum Z statistic generated was +0.76 and the minimum was -3.59. It seems that mixing the results from playing at different ranks will quite possibly impact the perceived results significantly.
So a big "attaboy" to yoyoma!
Dave Sigaty
"Short-lived are both the praiser and the praised, and rememberer and the remembered..."
- Marcus Aurelius; Meditations, VIII 21
Am I correct that ranking is only important as a guideline to getting decent games? Or that for some it is an endpoint and therefore a source of misplaced pride? Why not avoid the argument altogether by assigning an intuitive rank to yourself, and then not worrying about it? Unless you are a professional isnt focusing on rank similar to people arguing about who's house is worth more? Who cares? Better to look at it for feedback and then focus the energy doing problems, playing games, working on getting better no?
Obviously without ranking, you'd probably love to get a free even game with a very strong player who could help you improve your game. On the other hand, the very strong player might not be so amused to play a free even game with you when he is unable to determine just how weak you are.
http://tchan001.wordpress.com
A blog on Asian go books, go sightings, and interesting tidbits
Go is such a beautiful game.
