What's the value of endgame brawls due to better TB's use?

[S.Taylor]

OK, so we are seeing in TCEC that Stockfish has won 3 games vs houdini, but what is it worth to the average user, if these wins are only due to to SF's superior Endgame TB utilization, if the average user does not have those tb's on his computer?

e.g. If i want to have the strongest engine on my computer, paid or free, and i opt for SF because it wins more games so it must be stronger, but what if Houdini is actually stronger if it is without 7 men TB's? If so, isn't that a reason why it's worth paying for Houdini as being worth it because it is a stronger engine if it is minus TB's?

And, what do i get out of it, to just know that an engine KNOWS how to win due to its TB hits? What can a human learn from it?

[Ras]

only due to to SF's superior Endgame TB utilization

"Only"? That is a difficult problem to solve. Too little access, and your opponent steers the game towords a position that is won for him. Too much access, and it slows down your search so that your opponent wins in the middle game.

What can a human learn from it?

What can a human learn anyway from engines that strong? Making a certain move because 50 plies later in the game, there is a nice variation?

I'm using an engine for analysis only, and mainly for the eval graph. It it changes abruptly, that was a tactical error in the game. Otherwise, something positional.

And for validating lines in my opening book, of course, especially the sharp ones.

[Dirt]

...if the average user does not have those tb's on his computer?

Most people can get them on their computer. You can download them for free, or, if you have a slow internet connection, you can generate them yourself if you have a 64 bit computer with enough RAM.

...what if Houdini is actually stronger if it is without 7 men TB's?

Stockfish doesn't use 7 man TBs, and, as far as I know, Houdini doesn't either.

[syzygy]

OK, so we are seeing in TCEC that Stockfish has won 3 games vs houdini, but what is it worth to the average user, if these wins are only due to to SF's superior Endgame TB utilization, if the average user does not have those tb's on his computer?

Are you seriously saying that the average owner of a 44-core dual Xeon lacks the $40 necessary for buying a 120GB SSD?

e.g. If i want to have the strongest engine on my computer, paid or free, and i opt for SF because it wins more games so it must be stronger, but what if Houdini is actually stronger if it is without 7 men TB's? If so, isn't that a reason why it's worth paying for Houdini as being worth it because it is a stronger engine if it is minus TB's?

What does this have to do with TCEC?

And, what do i get out of it, to just know that an engine KNOWS how to win due to its TB hits? What can a human learn from it?

TCEC's stated purpose is not to show you what engine runs best on your hardware...

There are dozens of ranking lists for computer engines (and TCEC is not one of them). Just choose the one that you think fits best to the hardware you have and take the top-ranked engine.

I won't go into the substance of your "due to better TB's use" claim.

[Jouni]

Obviously Houdini 5 should set tb probe depth to 1 instead of default 5. SF8 has 1 and Komodo10 2 as default.