Has Top Engine Ever Lost a Game?

[jefk]

ah, mr viz is back

You played against a person who probably doesn't even use SF and maybe uses some old ass fritz or smth

yes, so what, as i said correspondence games can only be won at lower levels,
which i illustrated with an example

managed to "win" by playing stockfish moves

Can't you read ? i didn't use (mainly) SF moves, as -every move- i made an -in depth- analysis
with other 0usually more 'positional' top engines as (Obsidian, Beserk Lc0 etc).

as for 'achievement', well, do you play correspondence chess? Apparently not
because you didn't notice the originality of my opening move choices.
And you also can't be a really good human chess players otherwise you would notice the
intricacy of the practical endgame which resulted from this game in the end. While i'm
besides ICCF also a human -amateur chess player I'm still reading in this book
https://thinkerspublishing.com/product/ ... e-endgame/
And on chess.com already often manage to beat both human players (not always high Elo i admit) and intermediate 'bots'.
Yes you could also beat them cheating on chess.com with SF, that would be fun wouldn't it

Irrelevant ? well indeed, your nasty comments regarding my posting which i supposed could be
interesting for human chess players (rather than engine programming freaks) indeed were
complete irrelevant. Welcome to my foe/ignore list (just like a certain mosfel, btw)

Go back to your engine discord channel, troll, and stop bothering me.

[jefk]

as for the original question "has (a) top engine ever lost a game"

this depends of course on the time control, and opening(s).

For those interested in (top) engine tourns eg. in G.Banks (or playchess.com) style, the only way this can continue to be
interesting (besides unbalanced openings and double rounds as with TCEC )is imo with -much- faster time controls.
In other words, bullet, and lateron possibly with inc=0.

We know from the games of the illustrious mr Father that in such situations , time management
becomes important, and despited lots of progress in top engines (and lateron SF) in search during the
last 30 yrs or so, this problem still hasn't been solved, nor even improved, nor apparently even adequately
addressed. If mr Father can cause SF time forfeits, someone as Nakamura could also do that (eg in bullet 1-0)
and even with a lot more ease (eg with the 'father methods' as long as they still work in practice).

The -technical- problem isn't only time management, but accurately measuring remaining
time as allocated on the server, taking into account lag, and preferably knowing the time
situation of the (computer) opponent. Just like knowing the rating of the opponent,
also knowing exactly the amount of remaining milliseconds for the opponent only is possible
(I suppose) if such info is transmitted via Uci or otherwise; also some better more accurate
info about remaining time taking into account lag might be required from the server, so
there also might be some improvements needed in this area (now just speculating, i know).
But even without such knowledge, a perfect playing engine should be able to plan it's
moves within the amount of time in an accurate way, even if it would be a matter of
microseconds (hey, playing hyperbullet game with 0.1s each, that would be fun

Not so easy, egtb search takes time, you need to plan the 50 move draw rule, etc.
At least something still todo for engine programmers instead of endless tuning with
neural nets...(not that there isn't still some improvement possible, there imo also is).
my 2 cnts

PS and now leaving talkchess for a while (added to my blocksites list) coz i got a bit fedup with
some aggressive pricks here, not even knowing that mathematical proofs (and refutations,
and conjectures) are not a black and white thing (unlike eg. rules as at math olympiads)
but are a gradual process even with the discipline of pure math itself, as eg. known
since the work 'Proofs and Refutations' by the philosopher Imre Lakatos.
www.cambridge.org/core/books/proofs-and ... 5FBB9C6E10
https://en.wikipedia.org/wiki/Imre_Lakatos

[Viz]

I can read. And your claims are bogus.
"i didn't use (mainly) SF moves, as -every move- i made an -in depth- analysis
with other 0usually more 'positional' top engines as (Obsidian, Beserk Lc0 etc)."
And to surprise of literally no one what you achieved is playing sf mainline moves that it shows after 5 seconds of analysis, kek.
Sure it must feel so cool that you managed to use in depth analysis of "more positional engines" but as a result every single move you play can be retrieved from browser version of stockfish on lichess in 5-10 seconds, period - after like move 3.
Idk maybe for someone it's cool to beat people who probably don't even use engines while copypasting stockfish output, but I find no joy in doing so.

[shawn]

I can read.

But the mentally ill individual you are responding to can't

[chrisw]

I can read.

But the mentally ill individual you are responding to can't

One month ban.