TCEC S15, END of an ERA event is much more Brutal than I thought!

[nabildanial]

SF has a chance if LZ throughs away some games like it did last time.

So, that means no chance then because Leela has been very solid so far in this SuFi.

[Uri Blass]

After 26 games in TCEC SUFI, Leela is leading with 15-11.

My prediction for TCEC was just 55-45 Leela win.
cap4.PNG

Yes, I expected long ago 2:1 Win/Loss ratio for Leela, something on the lines +14 -7 =79. Let's see. I predicted that when the previous S14 was underway with 32930 net.

It IS an end of an era, if anything, the hardware for Leela is below the hardware for Stockfish in the current TCEC. GPU has much more room for improvement than CPU, we will assist in 1-2 years at a complete dominance strength-wise of NN-bases engines. FPGA might come next for chess engines, but it's a bit more involved to make them play chess.

I think that price comparison is wrong.
From my point of view if it is possible to generate a better chess player with lc0's hardware relative to Stockfish's hardware then lc0's hardware is better.

Edit:I understand that price comparison may be interesting for buyers but definition of the term better hardware cannot be dependent on the price.

[Laskos]

After 26 games in TCEC SUFI, Leela is leading with 15-11.

My prediction for TCEC was just 55-45 Leela win.
cap4.PNG

Yes, I expected long ago 2:1 Win/Loss ratio for Leela, something on the lines +14 -7 =79. Let's see. I predicted that when the previous S14 was underway with 32930 net.

It IS an end of an era, if anything, the hardware for Leela is below the hardware for Stockfish in the current TCEC. GPU has much more room for improvement than CPU, we will assist in 1-2 years at a complete dominance strength-wise of NN-bases engines. FPGA might come next for chess engines, but it's a bit more involved to make them play chess.

I think that price comparison is wrong.
From my point of view if it is possible to generate a better chess player with lc0's hardware relative to Stockfish's hardware then lc0's hardware is better.

So, you define "better hardware" as the hardware which allows for a better chess player on that hardware? I guess you will never be wrong with your assessment of hardware as defined by you.

[Uri Blass]

After 26 games in TCEC SUFI, Leela is leading with 15-11.

My prediction for TCEC was just 55-45 Leela win.
cap4.PNG

Yes, I expected long ago 2:1 Win/Loss ratio for Leela, something on the lines +14 -7 =79. Let's see. I predicted that when the previous S14 was underway with 32930 net.

It IS an end of an era, if anything, the hardware for Leela is below the hardware for Stockfish in the current TCEC. GPU has much more room for improvement than CPU, we will assist in 1-2 years at a complete dominance strength-wise of NN-bases engines. FPGA might come next for chess engines, but it's a bit more involved to make them play chess.

I think that price comparison is wrong.
From my point of view if it is possible to generate a better chess player with lc0's hardware relative to Stockfish's hardware then lc0's hardware is better.

So, you define "better hardware" as the hardware which allows for a better chess player on that hardware? I guess you will never be wrong with your assessment of hardware as defined by you.

The better hardware is not known in this case because I do not know what is possible to do with the hardwares.
It is certainly possible to produce something better than stockfish with the cpu and something better than lc0 with the gpu.

[Jouni]

OK now even I agree END of an ERA! Currently SuFi Win Probability LC0 96.7 % Stockfish 3.3% someone just calculated .

[Modern Times]

Whether Lc0 is stronger than Stockfish depends on how much an individual is prepared to spend on a GPU. That is the bottom line. For some it will be stronger, for others it won't.

I have a low cost GPU, so for me Stockfish is stronger by far.

[zullil]

I wonder if this defeat---assuming it happens---will spur any interest in (re)considering alternatives to "lazy SMP". Perhaps there really are better options, at least when core counts and time controls reach TCEC-levels.

[Laskos]

I wonder if this defeat---assuming it happens---will spur any interest in (re)considering alternatives to "lazy SMP". Perhaps there really are better options, at least when core counts and time controls reach TCEC-levels.

Lazy SMP is already very good on something like TCEC machine. On 43 cores the effective speed-up I guess is in the range of 20-25. With YBW it was maybe 10-12. I doubt another approach will go much above 30-32 with AB engines. It won't change much against Lc0. It's a bit the opposite, I guess Lc0 scaling to 4 and 8 GPUs can be considerably improved.

These are my guesses, don't take them at literal values.

[supersharp77]

After 26 games in TCEC SUFI, Leela is leading with 15-11.

My prediction for TCEC was just 55-45 Leela win.
cap4.PNG

Yes, I expected long ago 2:1 Win/Loss ratio for Leela, something on the lines +14 -7 =79. Let's see. I predicted that when the previous S14 was underway with 32930 net.

It IS an end of an era, if anything, the hardware for Leela is below the hardware for Stockfish in the current TCEC. GPU has much more room for improvement than CPU, we will assist in 1-2 years at a complete dominance strength-wise of NN-bases engines. FPGA might come next for chess engines, but it's a bit more involved to make them play chess.

These results are meaningless unless they are 100% Repeatable.....Anyone else been able to recreate LC0's (AlphaZero's) results vs Stockfish engines?

[mwyoung]

After 26 games in TCEC SUFI, Leela is leading with 15-11.

My prediction for TCEC was just 55-45 Leela win.
cap4.PNG

If you have been watching testing here. This is no surprise at all. Lc0 has been clearly stronger then stockfish for months when given very good hardware.