Carlos777 wrote: ↑Sat Feb 02, 2019 1:17 pm
A 12 games match is meaningless, you can't draw any conclusion from that.
Contests such as TCEC determine a champion.
They do not determine who is strongest (like CCRL / CEGT) but everyone knows that.
Still, it's wonderful chess on incredible hardware. Fabulous entertainment.
But we all know that.
Houdini is the deserving champion.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Carlos777 wrote: ↑Sat Feb 02, 2019 1:17 pm
A 12 games match is meaningless, you can't draw any conclusion from that.
Contests such as TCEC determine a champion.
They do not determine who is strongest (like CCRL / CEGT) but everyone knows that.
Still, it's wonderful chess on incredible hardware. Fabulous entertainment.
But we all know that.
Houdini is the deserving champion.
Don't forget that this was only the semifinal. Chances are good that Leela will win the final imo.
Yes, but that has no connection to my point.
Contests like TCEC do not show who is strongest.
They name a champion.
Which is not to say that the champion is not worthy.
All of the engines in the TCEC contest are a miracle of skilled, patient, meticulous effort by people of extraordinary mental acuity.
To win such a contest is a feat more amazing than a Usain Bolt sprint.
And that's pretty amazing.
Though the most amazing race I ever saw [by far] was by an Ethiopian woman who fell down in a middle distance race and got up for an incredible finish.
But my view of competition is probably not like most people.
Taking ideas is not a vice, it is a virtue. We have another word for this. It is called learning.
But sharing ideas is an even greater virtue. We have another word for this. It is called teaching.
Great job by Houdini and good to see that things are still competitive. Of course SF is probably the strongest engine but this is not testing the strongest engine but rather the winner in this tournament. If the strongest football team should always win the World Cup, Brazil would probably have double its Football World Championship Cups.
Dann Corbit wrote: ↑Sun Feb 03, 2019 7:52 am
Contests such as TCEC determine a champion.
They do not determine who is strongest (like CCRL / CEGT) but everyone knows that.
I don't know about that.
I don't see Leela in CEGT at all, and Leela appears as 122th‑125th place in CCRL, which is clearly wrong.
So CCRL / CEGT don't determine who is the strongest, either.
(I'm not claiming Leela is the strongest, I'm claiming that if Leela was the strongest CCRL / CEGT wouldn't determine it)
Your beliefs create your reality, so be careful what you wish for.
Easy, If you really would like to know the latest update of Leea elo, you can look at Leela discord, test channel info and also type !sheet. There are lots of sheets from testing at different time control and hardware setting from different testers.
But be aware, Leela developmemt is completely different from Stockfish. In Stockfish, 99% of their resources in development goes into testing, and in Leela 1% of development goes into testing and 99% goes into training.The main reason is that overview of development progress can be seen in Tensor Board( rather than testing individual networks, here and there)
Latest update, as 32930 is already selected for TCEC superfinal, developers encourages testers to invest their resources back into training, there will be fewer test results.
Sure, I just think that the title of this topic is too exaggereted.
I think this format is more susceptible to these kind of surprises than the regular format, because more games are played in the later. Good luck to Houdini in the finals.
Carlos777 wrote: ↑Sat Feb 02, 2019 1:17 pm
A 12 games match is meaningless, you can't draw any conclusion from that.
Contests such as TCEC determine a champion.
They do not determine who is strongest (like CCRL / CEGT) but everyone knows that.
Still, it's wonderful chess on incredible hardware. Fabulous entertainment.
But we all know that.
Houdini is the deserving champion.
.