eater.net/quaternions - I actually get quaternions now and how they're useful for rotations, and they're just amazing. Such elegance and simplicity, obscured by our minds' difficulty in visualizing and understanding 4D math and actions.
@El'endiaStarman yes, but remember dj's trying to optimize for nand gates not components. just optimize your individual components and you should be getting better results
It's a bit messy right now because I lost progress and started plugging in random things (and also I've been playing factorio, so I haven't been thinking about NAND gates for a good hour or 2)
Sometime I want to create a Universal Chess Opening System (TM). Basically... attempting to memorize all the various opening strategies is difficult, and not the best use of time for someone my level. So the question is... how can I get the maximum power : memory ratio for my opening strategy?
And while some "universal" systems exist, my approach will literally involve some sort of measurement of the complexity of the decision tree and memorization difficulty.
I wonder how these players stack up against computers. Not that computers will make mistakes, but rather are these players essentially perfect enough to force draws.
"...he [Magnus Carlsen] won't play it [his computer], because he just loses all the time and there's nothing more depressing than losing without even being in the game."
guess that answers it
Right now Caruana has the biggest lead anyone's had all game, with a 3/4th of a pawn advantage.
Unfortunately it's basically the endgame now and it's unlikely anyone will make any major mistakes... typically you need a ~3 pawn advantage or so to say that you are clearly winning.
In chess, one of the easiest ways to guess at who is winning is to look at how many and what types of pieces they have left. Empirically, these values are pawn=1, knight=bishop=3, rook=5, queen=9. In the computer age, this same scale has been adapted to provide an evaluation of the current board. These estimates I'm quoting include both material (# of pieces) and positional (where things are located) information.
It's a pretty... rough scale. But basically, if I've lost a knight (or, through lookahead, I'm destined to lose a knight), then my advantage drops by about 3.
These fractional numbers (like 0.88 in favor of black) are a result of a computer that's looked deep into the game tree and said that this position is worse than evenly matched but slightly better than most situations where white is down a whole pawn.
The big deal for the 3-pawn barrier is that, in the case of the very end of an endgame, a king + (a bishop or a knight) draws against a lone king. So it's not a matter of having the lead, it's about having enough of a lead that you can actually form a checkmating position from it.
You typically need a (1) two-piece lead, (2) a rook/queen, or (3) a pawn that's close to promotion to be in a winning endgame.
Right now it's Carlsen with 4 pawns and a bishop, against Caruana with 2 pawns, a bishop, and a knight.
Everyone knows that Caruana is "winning" (1 pawn ahead) but there's basically no way for either of them to end up with enough of a lead to actually win.
@betseg Nah, I think people are just optimistic. It's clear that both players really want to win, though, because it looks like they're playing this to the end.
@PhiNotPi yeah no, like Go, computers can still search way, way more possibilities than humans. it's why the best player in chess is still a human-computer hybrid iirc (although it's likely alphago has surpassed them?)
@PhiNotPi clearly you should pretend to be the world class player and think about your next move :P