Gamechanger
First among world champions Garry Kasparov reviews a book on computer chess and artificial intelligence as a pretext to explain it all in NYRB. It is as cogent a non-technical essay as any I've seen on what it means not only for chess, but for computing and beyond. Which is not to say that it's not quibblable, but that's not my aim here; instead, a few points on which to expand or expound:
Like so much else in our technology-rich and innovation-poor modern world, chess computing has fallen prey to incrementalism and the demands of the market. Brute-force programs play the best chess, so why bother with anything else? Why waste time and money experimenting with new and innovative ideas when we already know what works? Such thinking should horrify anyone worthy of the name of scientist, but it seems, tragically, to be the norm. Our best minds have gone into financial engineering instead of real engineering, with catastrophic results for both sectors.
Well, I'll quibble a tad here: Financial engineering (FE) has not deprived real engineering of real expertise; computer engineering has probably been more of a brain-drain on traditional disciplines. But what comprises financial engineering is largely the same kind of brute-force computation that chess programming relies upon. Unlike chess, finance is not a game of total information (efficient market theory doesn't go that far), and representations of the underlying processes and parameters are critical to the veracity of the outcome (whether for pricing or risk assessment: one common weakness of implementation was that the latter overrelied on the former, since pricing had to be more precise, on the "why waste time and money" principle). Model deficiencies were masked by the enrichment afforded by the Moore's Law expansion of computational power, whose better-faster-cheaper trifecta cached out other strategies (I'll forgo the martingale subreference here) before it broke the bank with Monte Carlo. (NB: Moore's Law is just as empirical [and to some degree as self-fulfilling] as many posited economic laws.) But the root deficiency was not in the models, but in how they were used within the banking establishment; what Kasparov says earlier about computer-assisted chess holds institutionally: Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.
Perhaps chess is the wrong game for the times. Poker is now everywhere, as amateurs dream of winning millions and being on television for playing a card game whose complexities can be detailed on a single piece of paper. But while chess is a 100 percent information game—both players are aware of all the data all the time—and therefore directly susceptible to computing power, poker has hidden cards and variable stakes, creating critical roles for chance, bluffing, and risk management.
Hidden behind this metaphor is the old saw about bipolar nuclear strategy (Soviet chess players, American poker players). (Hidden behind that is CIA misassessment of Soviet capabilites.) But, ex ante, I'll raise the point that it's still the wrong game: In pre-FE days, option modelers were chess-players, options traders were bridge-players ... and so's Warren Buffett, but then so's Bear Stearns' XCEO Cayne ... and of course there's Citi's XCEO Prince's penchant for Musical Chairs ... anyway, bridge mixes imperfect information into both competition and co-operation, not to mention its contractual aspects. But there is no right game, and that's why we keep playing; and, let's face it, looking for the optimal research algorithm is a mug's game, despite which it's one of the best ways to get ahead. (Game for the times? Yikes! I brought Mornington Nomic to the attention of David Chess ten years ago! Perhaps MCMC isn't Markov Chain Monte Carlo but Mornington Crescent Musical Chairs ...)
OK, more quibbling than I intended. But Read The Whole Thing, as it is written ... (also, I recommend Jonathan Schaeffer's book, One Jump Ahead, not mentioned in the article)
throwaway lines:
untitled
Kafka once met Einstein
they discussed the problem of our laws
unfortunately no transcript exists
related posts:
On chess: Calculus of Variations
Sample fiction: Inside Job
Like so much else in our technology-rich and innovation-poor modern world, chess computing has fallen prey to incrementalism and the demands of the market. Brute-force programs play the best chess, so why bother with anything else? Why waste time and money experimenting with new and innovative ideas when we already know what works? Such thinking should horrify anyone worthy of the name of scientist, but it seems, tragically, to be the norm. Our best minds have gone into financial engineering instead of real engineering, with catastrophic results for both sectors.
Well, I'll quibble a tad here: Financial engineering (FE) has not deprived real engineering of real expertise; computer engineering has probably been more of a brain-drain on traditional disciplines. But what comprises financial engineering is largely the same kind of brute-force computation that chess programming relies upon. Unlike chess, finance is not a game of total information (efficient market theory doesn't go that far), and representations of the underlying processes and parameters are critical to the veracity of the outcome (whether for pricing or risk assessment: one common weakness of implementation was that the latter overrelied on the former, since pricing had to be more precise, on the "why waste time and money" principle). Model deficiencies were masked by the enrichment afforded by the Moore's Law expansion of computational power, whose better-faster-cheaper trifecta cached out other strategies (I'll forgo the martingale subreference here) before it broke the bank with Monte Carlo. (NB: Moore's Law is just as empirical [and to some degree as self-fulfilling] as many posited economic laws.) But the root deficiency was not in the models, but in how they were used within the banking establishment; what Kasparov says earlier about computer-assisted chess holds institutionally: Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process.
Perhaps chess is the wrong game for the times. Poker is now everywhere, as amateurs dream of winning millions and being on television for playing a card game whose complexities can be detailed on a single piece of paper. But while chess is a 100 percent information game—both players are aware of all the data all the time—and therefore directly susceptible to computing power, poker has hidden cards and variable stakes, creating critical roles for chance, bluffing, and risk management.
Hidden behind this metaphor is the old saw about bipolar nuclear strategy (Soviet chess players, American poker players). (Hidden behind that is CIA misassessment of Soviet capabilites.) But, ex ante, I'll raise the point that it's still the wrong game: In pre-FE days, option modelers were chess-players, options traders were bridge-players ... and so's Warren Buffett, but then so's Bear Stearns' XCEO Cayne ... and of course there's Citi's XCEO Prince's penchant for Musical Chairs ... anyway, bridge mixes imperfect information into both competition and co-operation, not to mention its contractual aspects. But there is no right game, and that's why we keep playing; and, let's face it, looking for the optimal research algorithm is a mug's game, despite which it's one of the best ways to get ahead. (Game for the times? Yikes! I brought Mornington Nomic to the attention of David Chess ten years ago! Perhaps MCMC isn't Markov Chain Monte Carlo but Mornington Crescent Musical Chairs ...)
OK, more quibbling than I intended. But Read The Whole Thing, as it is written ... (also, I recommend Jonathan Schaeffer's book, One Jump Ahead, not mentioned in the article)
throwaway lines:
untitled
Kafka once met Einstein
they discussed the problem of our laws
unfortunately no transcript exists
related posts:
On chess: Calculus of Variations
Sample fiction: Inside Job
posted by nnyhav | 23:47
0 Comments:
Post a Comment
<< Home