I think time settings are also a decent sized factor when it comes to rank consistency.
A player might play much better with 30s byoyomi than 10s byoyomi, in that some mistakes might happen less frequently etc.
Or if you’re playing like 10 games in a row vs different opponents, you might be less fresh on your 9th game as the opponent who is just playing their first game today etc.
It would be interesting to see more of what you think about the ranking anyway, and if any of those kinds of factors might be relevant
Oh yes, for sure, the quality of my play varied over the course of a session (and between session), no doubt about it. But it’s not enough to explain the variation that I saw. There were plenty of games where I lost focus and blundered away the game against a lower-ranked player, or found a sharp move that turned the tables against a higher-ranked player. But I can tell when I’m straight up getting outplayed, or when I’ve got a clear edge on my opponent.
And the time settings don’t explain it either, for two reasons. First, the time settings I used were 5 min + 10 sec, so you can tell a player’s comfort level by comparing their play during the main time vs. byo-yomi. Some players took their time during main time, and then seemingly got caught off guard by byo-yomi and the quality of the moves dropped off. Others (like me!) pretty much blitzed the whole game, and the moves were pretty consistent throughout. I’m not spectacular at blitz, but I can hold my own, and so if someone like that crushes me then I’m pretty sure I’d have a tough game with any time control. (And vice-versa.)
Second, around half way through the event the variation was so sharp that I thought I was actually going crazy, so I performed an experiment with the Fox blitz pool. The games were much more consistent. I played on one account at my level and another account slightly below my level - my win rates where what you’d expect, but my wins on the first account and losses on the second account were all pretty hard-fought games, unlike how it went on OGS.
I can’t really quantify these observations in a way that could be measured at a large scale - I sort of tried last year, but it would take a lot more work. But the pattern was quite clear. I don’t care to speculate why the pattern exists - maybe there just aren’t enough dan players on OGS for the ratings to be fully accurate?
It’ll be good to find out eventually if it is a lack of dan players, something about the rating system itself like a deflation effect either from new accounts or other reasons.
I had a similar experience. I am not used to the time setting also. Another thing I noticed was a lot more preference for 3-4 openings from others. Perhaps everyone is sick of 3-3 invasions and having to play with influence? I pulled up data from Got Stats? to compare the 121 games from the challenge to the 230 games (5m+5x30s) before the challenge (which are a mix of settings, a lot of them longer).
Wins changed from 53% resign to 58% resign
Losses changed from 63% Resign to 41% Resign
Black W%/L% went from 39/61 to 47/53
White W%/L% went from 48/52 to 23/77
Funny to see Black and White flip! I seem to give up less while my opponents give up more. I think I didn’t want to end games too soon. I haven’t changed my openings preferences.
I do think there could be a deflationary effect at play. Some opponents a few stones weaker seemed a lot stronger than back then while I could also keep up with some opponents closer to my rating before the event. There needs to be more players because this website has a wonderful modern ui/ux. In any case, I had fun overall, which is the important thing.
I think the reason might be that the rating on OGS is quite volatile.
As I understand it, this is to get new players quickly at their level.
The drawback is that the rating of experienced players is shaky.
Another reason might be the mix of the rating with different board sizes.
I for instance sometimes play 9x9 games against lower ranked players and don’t put much effort in the game. So I lose some of these and my rating dropes significantly.
On the other hand I might play a 19x19 games which I take more seriously.
This could lead to my opponent feel surprised about the strength as pwsiegel experienced it.
It’s been stopping at random points but I get @ on Discord so I just restart it when it happens. It doesn’t have any effect on the results at all though.
90 hours left going by utc
Is doing 100 games possible?
Given: 5m + 7s and 5m + 5 x 30s, which takes about 25 mins (source: ogs default live time setting)
So three sleep periods of 8 hrs takes out 24 hrs and four takes 32 hrs. While one could opt for less sleep, they still gotta eat and do things lol.
If the average time is 25 mins with no matchmaking time, this would take about 42 hrs leaving 48 hrs for everything else. If there is 5 mins of matchmaking time average, it’s gonna be 50 hrs playing and 40 hrs else.
Does it have to be 100 games in a single board size, or can it be 100 games across the three?
Even if it doesn’t count to do 100 across three sizes, i’ll still be proud of myself for hitting the milestone. Most Go i’ve ever played in a single month already.
That’s pretty much what online hold’em players do lol. Seems too stressful with 30s byo, maybe at 1m. Correspondence would take forever if the opponents aren’t prompt. Having multiple monitors would also help…
