Discussion in 'Politics & Current Events' started by Q*bert Jones III, Sep 15, 2022.
You make Joseph Saddler cry.
That would also be a creative way to develop more interest in chess among the youth.
So combining my three great internet pursuits - watching people create electronic projects and watching people play chess - this video shows one person's attempt to find out if it is actually possible to build the rumored device and how well it works.
I really don't want to click on that link, so I have to ask... Does it work?
No, not clicking. Lemme know what he tells you
I watched the video. It was entertaining. Summary: it worked. The first half is the guy building the thing, the second half is two games against a good chess player, first without the cheating device, then with the device. The good player could tell the guy was cheating but couldn't figure out how. (Spoiler: the device was strapped to the guy's leg.) Conclusion: it would be possible to cheat using a device like this if the device can be concealed - the method of concealment doesn't necessarily need to be the same way Christopher Walken kept Bruce Willis' dad's watch safe in Pulp Fiction.
Maybe he should have gotten "Cheaters Justice."
Leg device has been used before IIRC. It's not really much different to cheating at cards
I'm still trying to figure out how you cheat at chess. It's not like he's got an accomplice looking at his opponent's cards in a poker game.
The cheating is done by searching a database which contains millions of games - a collection of all of the moves in those games and the results of those games. You input the current chessboard and the database finds all the games where that exact same board happened, and which move resulted in the most wins. It's not like Deep Blue is playing by analyzing the board and all the possible moves, it's just searching an extensive database to find which move gave the best results in similar situations in the past.
The good player in the video above, and the chess grand master who accused the original guy of cheating which started this thread, both said the same thing - the strategy employed by the cheater didn't make any sense, they weren't building up a plan and then executing it, and they found it frustrating because they're used to playing non-cheaters who build up a strategy which you can then counter with your own strategy. The cheaters were doing highly unexpected things which was throwing the good players off their game.
Okay -- I guess that makes sense. But it just seems that without a strategy, the cheater is just making single moves based on what the board looks like at each move -- intuitively that seems like something that wouldn't be all that successful. But I'll take the Grandmaster's word for it.
It's also that the cheating "non-strategy" lines up almost exactly with the chess engines. Great players can do this every once-in-a-while. Cheating cheater was making 90%+ "best moves" in a game over several tournaments at a rate like 3-4x as often than Magnus at his absolute best ever. So, this kid either caught lightning in a bottle for a year, who also happens to have been caught cheating before and admitted to cheating before.
OR....... and hear me out.......
he's a cheating cheater and just managed not to get caught during the act this time.
There are basically 2 kinds of assistance
There are the table bases which contain vast amounts of opening preparation and analysis. In my day this stuff used to be in a book but now is all computerised. Grandmaster play is a lot about your preparation - i.e preparing opening lines and learning them. Getting people "out of prep' is a tactical advantage because then they have to work out the moves over the board. Top games can go 10-20 moves blitzed out because the players know the lines by rote. There are also end game table bases i.e positions that have been solved. You need to know all of these. So knowing the prep is a big advantage to a cheat as you can play all best moves in the opening.
Then there are the engines that simply crunch the positions and calculate the best moves.
The last generation of engines used both table bases and crunch power.
As I understand it, the latest engines like Alphazero use something more like true AI so they don't need table bases.