Is percentage chess good?

[qpawn]

Is it a good idea to play chess in percentages? For instance someone told me that in endings between a rook vs two bishops the bishops win 75 % of the time.

This could be an improvement over the...um...guesswork of certain inept chess players who start threads on chess. Of course I am not mentioning anyone.

Suppose a player had a reliable list of endings for transposition. If the percentage of open board endings with a bishop better than a knight was, say, 60 % [I made that figure up] then the player could transpose into the ending with the bishop and be at least 60 % sure of, at worst, equality.

I have come across no book that lists endings for transposition with thses percentages. Is there any such list? For such an ancient game as chess I am astonished nif someone has never tried to compile one.

[Kevin Bonham]

The thing you have to watch in going by these rules of thumb blindly is that a very experienced player will know that there are certain types of positions that go against the trend. I know that if I am thinking about going into a bishop-vs-knight ending I'm not thinking "is a bishop better than a knight 60% of the time?". I'm thinking "is this one of those positions where a bishop is better?" If I can answer "yes" to that question then it doesn't really matter whether the bishop is better in 30%, 50% or 70% of positions. In fact I've never learnt the answer to that question; I've only tried to learn what type of positions each piece is better in. Knowing the general patterns is one thing but you become a better player if you also know when to break them.

Something that is worth knowing is that there are some kinds of positions where theory says one side is better but real-game experience says something else. Most players, brought up on the 9-5-3-3-1 scale of piece values, would prefer to have two rooks plus pawns against a queen plus the same number of pawns in the ending. Yet there was a statistical survey which showed that the side with the queen scored 52% in such positions. It is simply easier to use the queen in practical play for most players.

I would be interested to see some stats on exchange up compared to pawn up. I actually prefer being the latter in most positions.

[qpawn]

Yes, I think you make some valid points.

There are exceptions to rules of thumb. One occurs in my game against Fritz that is in the other thread in chess training. I exchanged off into an ending of bishop against knight with equal material of a few pawns and a rook each. I thought that I would be at least equal with the bishop. But Fritz gave this as a bad decision: I should have kept some more material on the board for a different late middlegame/endgame in which I would have had a slight plus.

Then I lost the game becuase I had a choice between bringing out a bishop or my king. I chose the bishop and brought out the king next move. Wrong decision.

[FM_Bill]

I'm not too keen on percentage chess as it seems too general and is a distraction from details. Having a white pawn on a2 or a3 might make a critical difference.

You can increase the *probability* of things going your way by small factors. For example, if a pawmn is protected, you are less likely to lose it than if its not.

On being the exchange up, Capa once said you can often give back the exchange to be a clear pawn up. The exchange is really iffy as it depends SOOO much on the position.

The bottom line is who will have the initiative in the long term?

[likesforests]

Hmm. Imagine you only knew "Rook endings are drawn 52% of the time" but you didn't know Philidor, Lucena, the short-and-long defense, the frontal attack, Vancura, Cheron, Kopaev, rooks to open files, rooks belong behind passed pawns, activity is often worth a pawn, Flohr-Vidmar, etc. How well would you fare against an opponent who has a basic knowledge of such endings?

I once decided to ignore minor piece endings and get by on maxims like:

* Knight endgames are pawn endgames and knights hate rook pawns.

The next week I reached a N+3P vs N+2P endgame. I sacrificed my 7th-rank central passed pawn to reach an N+2P vs N endgame where I had two wing pawns on the 3rd rank. Knight endings are pawn endings, right? It turned out that maneuver was a terrible blunder drawing the game. That motivated me to finally learn something about such endings. I think knowing 5 positions is worth knowing 25 maxims.

[Capablanca-Fan]

Hmm. Imagine you only knew "Rook endings are drawn 52% of the time" but you didn't know Philidor, Lucena, the short-and-long defense, the frontal attack, Vancura, Cheron, Kopaev, rooks to open files, rooks belong behind passed pawns, activity is often worth a pawn, Flohr-Vidmar, etc. How well would you fare against an opponent who has a basic knowledge of such endings?

Absolutely hopelessly. And if you know all the above, you're a much better R endgame player than most in the Aussie chess scene.

Compare what KB said above. Note also, according to Watson, the top players don't even care about rules but what is good in a particular position. (My only gripe is that he thinks this is a modern development, but I found examples of this point in Tarrasch).

I think knowing 5 positions is worth knowing 25 maxims.

Also good. E.g. it is useless knowing that KP v K or KRP v KR is winning in X% of cases. You must know which specific cases.

[Davidflude]

The thing you have to watch in going by these rules of thumb blindly is that a very experienced player will know that there are certain types of positions that go against the trend. I know that if I am thinking about going into a bishop-vs-knight ending I'm not thinking "is a bishop better than a knight 60% of the time?". I'm thinking "is this one of those positions where a bishop is better?" If I can answer "yes" to that question then it doesn't really matter whether the bishop is better in 30%, 50% or 70% of positions. In fact I've never learnt the answer to that question; I've only tried to learn what type of positions each piece is better in. Knowing the general patterns is one thing but you become a better player if you also know when to break them.

Something that is worth knowing is that there are some kinds of positions where theory says one side is better but real-game experience says something else. Most players, brought up on the 9-5-3-3-1 scale of piece values, would prefer to have two rooks plus pawns against a queen plus the same number of pawns in the ending. Yet there was a statistical survey which showed that the side with the queen scored 52% in such positions. It is simply easier to use the queen in practical play for most players.

I would be interested to see some stats on exchange up compared to pawn up. I actually prefer being the latter in most positions.

this post raises lots of good points.

First many Europeans use a different count

pawn = 1
Knight = 3
bishop = 3
rook = 4.5
queen = 8.5

Never the less the position is still the main thing. also cobinations of pieces has a big impact.

With no other pieces on the board

queen + knight is often much better than queen and bishop.

queens + bishops of opposite colours is any thing but drawish.

Playing the exchange ahead a common technique is to give back the exchange winning a pawn in the process. This is why winning the exchange at the cost of a pawn often leads to difficulties in winning the game. Because if you give back the exchange to win a pawn the pawns wind up equal.

I have just drawn a game a correspondence game against a player with an ICCF rating of over 2400 where I was an exchange and a pawn down but there were compensating factors. (I consider this to have been the toughest correspondence game that I have ever played and it was no easier for my opponent. He took four weeks plus two weeks holiday on one move). I am currently annotating the game and even that is anything but easy.

[Aaron Guthrie]

Only some of the time.

[Garvinator]

Only some of the time.

I would think it depends on the percentages.

[Southpaw Jim]

I would think it depends on the percentages.

But which ones?

[Adamski]

IFF the percentage of games I win is good, THEN percentage chess is good!

[Aaron Guthrie]

But which ones?

Not all of them.

[Javier Gil]

this post raises lots of good points.

First many Europeans use a different count

pawn = 1
Knight = 3
bishop = 3
rook = 4.5
queen = 8.5

Which Europeans?

I'd give the knight = 2.9 (it clearly is a little worse than a bishop in a larger percentage of positions)
Rook = 5
Queen = 9.5 (definitely more than 9 and less than 10)

I guess everyone has his own nowadays.

[Basil]

I'd give the knight = 2.9 (it clearly is a little worse than a bishop in a larger percentage of positions

Huuuuuuuuuuuuuuuuge call!

[Adamski]

I hope you resisted the urge to have a smoke while typing all those "u"'s, Gunner. All the best with your endeavours to give up the weed. Lots of us are right behind you in this.