Statistics based chess engine?

I was wondering if anyone knows of any chess engines that are purely based off of analyzing a large repetoire of high level matches and going through the win rate of each move played in a situation.

I don't know but I don't think that would be extremely useful. Many games are new within 10-12 moves, often sooner, with the exception of some really booked up people and high-level players. What I mean by that is after 10-12 moves, no games would exist in the database that match the current game being analyzed.

So they would quickly run out of valid positions to match against.

Can you explain exactly what the engine is doing in that picture because your post doesn't appear make any sense. It sounds like you're describing a calculator.

The Chessbase app for Android does that.

Does what? What is he describing?

For an example, I have an OTB tourney game played by a Class A/Expert player and me. By move 9 by white, we are out of my database from TWIC and by move 11 by black. That is pretty normal for my games.

He is basically asking if there is an engine that works kind of like how some people play correspondence .

Look at the database, pick the move that gives the highest winning percentage for the side to move.

The OP is asking if any chess engines show the win rate of each move made/  (That is my assumpton.)  My reply is based off of that.  The chessbase app can be configured to show engine moves by popularity, or by win rate.

I'm obviously being thick so I apologise. How is what you/the OP are/is describing different to simply looking at a database? What is the engine in question here doing to further analyse the position?

I don't think there is a difference. He was just wondering if any engines use database lookups/statistics to determine moves.

Sorry Diakonia, I still don't get it :(. I'm obviously badly misundestanding something here.

Oooh. He means using an engine to play a game, not to analyse a position? That makes sense. That never occurred to me. Sorry about that. Thanks.

Or im not splaining it correctly...

Highlighted the terms, off to find one. If there aren't any, there will be soon.

The problem is that the number of chess positions is HUGE. Even with the largest databases, including all the chess playing sites in existence along with all OTB games ever played, the engine that uses database stats only, will get to a position that has never been played before. How is it going to make that next move?

And that large of a database includes a bunch of junk, poorly played games. If you restrict it to a subset of good games, you are more likely to reach a unique position much more quickly.

Would that even be called an engine ?

Sounds more like a simple script selecting the statistically most successful move to me.

so what happens when a line of play that historically wins a lot is refuted or equalized in a high level game and suddenly stops getting played? Aren't the statistics wrong in that case?

An idea is giraffe: https://bitbucket.org/waterreaction/giraffe

Have you given http://www.365chess.com/opening.php a try?

I'm surprised no one has mentioned chess.com's Game Explorer yet. It can't select the statistically most successful move automatically, but it shows the stats for each move. You just need to do little bit of simple calculation, and then click on the desired move yourself.