Opening 32Z: reply errors
Posted By: Timothy Chow In Response To: Opening 32Z: reply errors (Nack Ballard)
Date: Saturday, 2 January 2010, at 3:13 p.m.
In Response To: Opening 32Z: reply errors (Nack Ballard)
Nack Ballard wrote:
Or, if they induce equivalent overall errors, why not just make the play with the higher equity (32S)?
The point is that it's totally unclear whether 32S would end up with higher equity if I and my opponent were to play 46k games with each other. One can't just blithely assume that if my opponent and I are roughly equal then the equities will be the same as with bot-vs-bot. Snowie is roughly equal to itself and GNU is roughly equal to itself; does that mean that Snowie-vs-Snowie and GNU-vs-GNU equities can be assumed to be equal to each other?
The plays are close enough that the decision is, or should be, outweighed by other considerations.
History tends to remember the reversals (they're more interesting) a lot more than the gaps that widen.
Do you have a reason to believe that the current [S Z14.5] perception of value is more likely to invert to [Z S5.5] than to widen to [S Z34.5]?
No. But that's not relevant to my point, which is that enslaving oneself to fashion carries its own risks, of inhibiting the flexibility of one's mind, and falling into the trap of thinking that "current theory = absolute truth." If you look at my website, you'll see that I am always careful not to say that a play is "correct" just because that's what a bot thinks is correct. I always refer to "current theory" or I say that GNU "disagrees with" Robertie. I'll only use absolute language if I'm really convinced we've hit absolute truth.
Compare with chess. I believe it's a mistake, at all levels, to play only the "theoretically best" lines. At the top levels, of course, a large part of opening preparation amounts to finding lines that refute current theory. But even if you're not at that level, it's important not to fall into the trap of playing by rote. It's also important to expose oneself to a wider range of positions to gain more experience. Of course there's a limit here; in an important game, you don't want to play a line with an obvious and crushing refutation. Within limits, though, I strongly believe it's more valuable to test the boundaries than to adhere slavishly to current theory.
Messages In This Thread
BGonline.org Forums is maintained by Stick with WebBBS 5.12.