I find it hilarious how the vid starts with Eric making a joke about how the commenter spelt bongcloud, only to spell it wrong himself in the video title
@@Jason-sq2up The developers were playing around with Mittens' level, so it was at some points far below its usual strength. That explains Nemo's win for sure and probably Ben's as well. Bok just played amazingly though
Jonathan Schrantz has beaten SF without help before, there are lines in the Stafford where you can convince it that its winning when you actually basically have a fortress. The same thing should be possible against Mittens
"I'm a cat toy that can't be eaten" - Eric Rosen 🤣 . Also , the prophecy must not be fulfilled probably refers to Eric not wanting to play a 150 move game . The game lasted 130 moves .
Yes Eric this is what we wanted to see Bonk the damn kitten on his head of the Anger that people have after getting their soul being taken as they started the match thinking 1 elo
Very similar to the game I played against Mittens. I played the first 15 moves entirely on my own (no dodgy openings, just played it like a normal game) then let Stockfish take over my terrible position. Eval was about -3 for Mittens at this point, and SF got it back to 0.0 Then had about 100 moves of shuffling back and forth and trading some stuff in a completely drawn position, and eventually drew by repetition in KP v K endgame after 146 moves. Tedious playing it out, but interesting. Seems like SF can overcome a 2.5 to 3 eval disadvantage to get a draw.
I took a class once on AI for chess engines, and Mittens' refusal to accept a draw in the endgame here reminds me of the wonderfully-named "Contempt Factor", where you artificially decrease the evaluation of a draw against opponents you expect to be vastly superior to (such as human players), in the hopes that your opponent will play sub-optimally and screw up the drawn game. Without a contempt factor, the AI is usually fairly willing to accept a draw if it thinks it's inevitable (imagining it playing against itself). I suspect that that's what Mittens is doing here, but I'm not certain of what's standard in these modern engines. For example, if Mittens is using an endgame tablebase, it knows for a fact that the game is absolutely drawn from a mile away. But I'm not sure if using endgame tablebases for human-facing bots is considered too rude. One subject that class never covered was how to make our chess engines fun to play against, rather than as strong as possible.
Mittens does use a endgame tablebase, but that's not relevant here. Tablebases are for perfect play. Mittens is playing against a human who doesn't have a memorized tablebase. If Mittens has a high contempt factor, it's going to extend the game, hoping for the human to blunder the rook.
It basically makes it a case of 'Can Stockfish beat Mittens with two move odds and without casting rights?'. Interesting, but it would have been more fun to try and see if Stockfish could hold on with the king on e2 for longer.
I don’t know the opening but a while back I saw penguin doing something like. Pawn G3. Knight h3 eventually to F2 the F pawn somehow pushed to F4. And the light square bishop got to G2.
holy shit this thumbnail hahahahahah omg I haven't even watched the video yet. I clicked the link solely and exclusively because of that incredible thumbnail image.
Can we see your thoughts on something like a Breyer or Chigorin? A nice, classical Spanish with your commentary would be excellent. I really miss your non-dubious days
My first day starting chess and I'm very confused. A machine kitten playing another robot and a bongcloud is involved. If a human, press 1. bots, don't read that!!!!
Please do not take this "suggestion" seriously, but if you're going to continue using Stockfish against Mittens, it might be funny to see g4 or the Tumbleweed on Move 2 of Bird's Opening. -I just played both suggestions against Mittens without any engine help, and of course got obliterated (in single-digit moves during the g4 suggestion, but I did make it to Move 25 in the Tumbleweed).-
Hey Eric! Just wanted to let you know that using MultiPV (having SF show multiple lines at once) causes it to lose elo. So if you want it at full strength, make it show only one line.
I’m probably wrong but in all the Hikaru videos I’ve watched (many), 3. Ke1 was never played. In other words it’s not really a bomgcloud or bungloud or bumproud or etc.
