61–64 of 64 posts
I play Go a lot, and chess occasionally. I think it's about mindset not skillset - games like these are about solving local problems ('tactics') to achieve a greater aim ('strategy'); people who are attracted to them are generally people who like the sort of problem solving that comes up in programming, hence the cross-over. Your question, however was whether or not the skills translate - I don't think that solving d…
neil, are you on OGS?
Earlier quoted context omitted.
I've never understood the 'imperfect information' thing when discussing chess. Chess _as played by people_ is full of hunches, uncertainty, prejudice, intimidation. Just look at how Kasparov played or Lasker, to see how important psychology is to chess. Or go down to Washington Square Gardens and play under the glare of dozens of people tutting at your every move. It's so much more than an abstract problem.
Yes, but the possible input to a chess game is strictly limited - one player cannot just change the way a piece moves or introduce a new piece, and thus, it is not taken into account in any strategy. These sorts things must be taken into account when programming.
Earlier quoted context omitted.
The original quote was: "Where it might frustrate a chess player, is that chess lacks an element of imperfect information." Any chess player worth their salt deals with imperfect information, and has to manage risk, make kludges, pick imperfect paths, deal with the clock etc etc.
Right, however none of that has to do with perfect/imperfect information. To quote Wikipedia: "In game theory, an extensive-form game has perfect information if each player, when making any decision, is perfectly informed of all the events that have previously occurred." This is a technical designation, it has nothing to do with the future, only the past and current state. Yes, to be a good chess player you need to b…
In this context the fact that chess has perfect information in a technical sense is irrelevant, since chess as played is so much more than a game about digital information.
From my experience in coding as well as hiring coders, I've actually found that the best programmers are those that can make deductive arguments quickly. Most argumentative people argue inductively with anecdotes, references to studies, analogies, etc. But when you find someone who argues deductively by slipping in premises for tacit or explicit agreement and then lowering the boom with an axiomatic conclusion, this person usually creates good code for me.
For computers, chess games are deductive problems where all (or almost all) possible moves (arguments) are examined to choose the best one, which is only made possible by binary brains. For humans, chess is extremely inductive - analogies, patterns, general underlying strategies albeit with changing particulars, etc. The best human chess players deductively analyze multiple possibilities of these inductive patterns (read Bobby Fischer - who I have not played, Tal Shaked - who I have played, and others on the amazing way they do this). So while the skill is not transferable, there are some similar aptitudes in good programmers and good chess players.
I know there's a holy grail out there of the ultimate programming skills assessment test, and I don't have it. But what I do is this: I give two programming problems (no language or syntax required at all, just the problem basically in comments). They have to solve it in a robust way (and every now and then someone comes up with a brilliant solution I haven't seen yet, although most people suggest the same stuff over and over again), and then they have to defend why the solution is good while I time them.
Most people can learn to think (and argue) deductively, but very, very few that I have found do this instinctively (and can thereby do it 10x faster than the average person). When I find someone who can even do it 2x faster than the average bear, I hire that person. It's just so rare.
I know that's not a perfect test, but as someone who both plays chess and also is constantly looking for a good indicators for programming skills, this is what I do. I've had pretty decent success this way. So who knows, according to my methods, maybe the best lawyers and philosphers are actually the best programmers.