I think the leela fans prefer "leela was disconnected", which if it is the case may allow the game to be replayed according to the rules.chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
TCEC Division 3 results simulator
Moderators: hgm, Rebel, chrisw
-
- Posts: 228
- Joined: Sun Mar 12, 2006 3:11 pm
Re: TCEC Division 3 results simulator
-
- Posts: 1346
- Joined: Sat Apr 19, 2014 1:47 pm
Re: TCEC Division 3 results simulator
Yes. But it is two big if.
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: TCEC Division 3 results simulator
Now you see that the gut feeling of an enlightened person like Kai is more stable than Sim wild oscillations, which are a sign of a flaw, at least until the last several games. It's hard to comprehend oscillations from 3% to 30% of Sim with many games remaining, while Kai's 10% to 20% oscillations are comprehensible by an educated person .chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
-
- Posts: 4611
- Joined: Wed Oct 01, 2008 6:33 am
- Location: Regensburg, Germany
- Full name: Guenther Simon
Re: TCEC Division 3 results simulator
I want to add that also Hannibal 'disconnected' just a few games before this game.
No one questions though, if it really was its fault. (missing fan hype?)
May be Edsel can say something on this matter after he had seen the logs.
https://rwbc-chess.de
trollwatch:
Talkchess nowadays is a joke - it is full of trolls/idiots/people stuck in the pleistocene > 80% of the posts fall into this category...
trollwatch:
Talkchess nowadays is a joke - it is full of trolls/idiots/people stuck in the pleistocene > 80% of the posts fall into this category...
-
- Posts: 4321
- Joined: Tue Apr 03, 2012 4:28 pm
Re: TCEC Division 3 results simulator
hahaha!! I diagnosed my Sim as manic depressive alreadyLaskos wrote: ↑Thu Aug 16, 2018 11:42 amNow you see that the gut feeling of an enlightened person like Kai is more stable than Sim wild oscillations, which are a sign of a flaw, at least until the last several games. It's hard to comprehend oscillations from 3% to 30% of Sim with many games remaining, while Kai's 10% to 20% oscillations are comprehensible by an educated person .chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: TCEC Division 3 results simulator
chrisw wrote: ↑Thu Aug 16, 2018 11:49 amhahaha!! I diagnosed my Sim as manic depressive alreadyLaskos wrote: ↑Thu Aug 16, 2018 11:42 amNow you see that the gut feeling of an enlightened person like Kai is more stable than Sim wild oscillations, which are a sign of a flaw, at least until the last several games. It's hard to comprehend oscillations from 3% to 30% of Sim with many games remaining, while Kai's 10% to 20% oscillations are comprehensible by an educated person .chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
-
- Posts: 4321
- Joined: Tue Apr 03, 2012 4:28 pm
Re: TCEC Division 3 results simulator
I cut the tournament rating adjuster back a bit, but it doesn't make much difference. At this stage, played 90, 22 to go, the manic changes in prediction by results so far are going to happen. "Games so far" predominate.Laskos wrote: ↑Thu Aug 16, 2018 11:54 amchrisw wrote: ↑Thu Aug 16, 2018 11:49 amhahaha!! I diagnosed my Sim as manic depressive alreadyLaskos wrote: ↑Thu Aug 16, 2018 11:42 amNow you see that the gut feeling of an enlightened person like Kai is more stable than Sim wild oscillations, which are a sign of a flaw, at least until the last several games. It's hard to comprehend oscillations from 3% to 30% of Sim with many games remaining, while Kai's 10% to 20% oscillations are comprehensible by an educated person .chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
Each game in a sim is predicted by elo, and the "best" way to produce a Kai gut prediction of higher LC0 chances is simply to increase the LC0 initial elo estimate.
The other reason for the apparent manic behaves is the performance of Ethereal. That's at top spot, and rightly gets almost 100% chance of first place, and 0% chance of second. So the rest of the pack are all competing for a share of 100 pts, whereas in a field with no clear outright winner, their score would be their share of their 1st place chance plus their second place chance, which a) has the effect of increased "sharing" out of pts by the addition process and b) more pts to share. Those effects tend to spread the range and increase the numbers.
Sim accounts for everything (that I can think of, which may be broken of course)
-
- Posts: 4321
- Joined: Tue Apr 03, 2012 4:28 pm
Re: TCEC Division 3 results simulator
Situation after Arasan just beat DX (with LC0 disconnect/server crash included)
Arasan 76%
Pedone 15%
LC0 5%
And same situation, but with LCO given the replay, and winning
Arasan 66%
Pedone 11%
LC0 21%
Arasan 76%
Pedone 15%
LC0 5%
Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two
Ethereal 3315 3300 1.000 1.000
Arasan 3150 3124 0.000 0.764
Pedone 3117 3092 0.000 0.158
lc0 3116 3100 0.000 0.055
Hannibal 3140 3158 0.000 0.014
DeusX 3120 3130 0.000 0.009
Nemorino 3092 3099 0.000 0.000
Bobcat 2982 3029 0.000 0.000
Arasan 66%
Pedone 11%
LC0 21%
Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two
Ethereal 3315 3300 1.000 1.000
Arasan 3150 3124 0.000 0.664
lc0 3128 3100 0.000 0.214
Pedone 3117 3092 0.000 0.116
DeusX 3120 3130 0.000 0.005
Hannibal 3127 3158 0.000 0.000
Nemorino 3092 3099 0.000 0.000
Bobcat 2982 3029 0.000 0.000
-
- Posts: 10948
- Joined: Wed Jul 26, 2006 10:21 pm
- Full name: Kai Laskos
Re: TCEC Division 3 results simulator
Yes, with only 5 games to play and one eligible place, oscillations are inevitable by now. Arasan just beat Deus, so Deus is out by now (1%). Lc0 still has some chances, but IMO no higher than say 8%. That is, if no replay of that disconnected for unknown reasons game of Lc0.chrisw wrote: ↑Thu Aug 16, 2018 12:32 pmI cut the tournament rating adjuster back a bit, but it doesn't make much difference. At this stage, played 90, 22 to go, the manic changes in prediction by results so far are going to happen. "Games so far" predominate.Laskos wrote: ↑Thu Aug 16, 2018 11:54 amchrisw wrote: ↑Thu Aug 16, 2018 11:49 amhahaha!! I diagnosed my Sim as manic depressive alreadyLaskos wrote: ↑Thu Aug 16, 2018 11:42 amNow you see that the gut feeling of an enlightened person like Kai is more stable than Sim wild oscillations, which are a sign of a flaw, at least until the last several games. It's hard to comprehend oscillations from 3% to 30% of Sim with many games remaining, while Kai's 10% to 20% oscillations are comprehensible by an educated person .chrisw wrote: ↑Thu Aug 16, 2018 11:18 amOh wow, did LC0 disconnect? That would indeed changes things if replayed, would be temporarily in second place in cross table.
Affects the Sim prediction, LC0 jumps to 28% promotion chance
90 games played ...Code: Select all
Engine, Tournament Elo, Initial Elo, First, First two Ethereal 3333 3300 1.000 1.000 Arasan 3169 3124 0.000 0.494 lc0 3166 3100 0.000 0.287 Pedone 3149 3092 0.000 0.185 DeusX 3121 3130 0.000 0.034 Hannibal 3089 3158 0.000 0.000 Nemorino 3083 3099 0.000 0.000 Bobcat 2919 3029 0.000 0.000
Each game in a sim is predicted by elo, and the "best" way to produce a Kai gut prediction of higher LC0 chances is simply to increase the LC0 initial elo estimate.
The other reason for the apparent manic behaves is the performance of Ethereal. That's at top spot, and rightly gets almost 100% chance of first place, and 0% chance of second. So the rest of the pack are all competing for a share of 100 pts, whereas in a field with no clear outright winner, their score would be their share of their 1st place chance plus their second place chance, which a) has the effect of increased "sharing" out of pts by the addition process and b) more pts to share. Those effects tend to spread the range and increase the numbers.
Sim accounts for everything (that I can think of, which may be broken of course)
-
- Posts: 4321
- Joined: Tue Apr 03, 2012 4:28 pm
Re: TCEC Division 3 results simulator
looks like they will call it a server disconnect, so - replay game presumably.
Do they play it now, or at the end?
Do they play it now, or at the end?