Handicapping is routine in the Japanese game Go (my best game). The basic strength unit is one stone, and a one-stone difference represents a full level of difference in strength.

I (about a 1500 player) once asked a 2100 player how much of a handicap she would need to give me so that we would have an equal chance to win. "Probably a knight, maybe more," she answered. I once took a knight handicap against a 2200 player and lost, but it was a much tighter, closer game than one with no handicap.

That might suggest that a pawn is equivalent to about 200 points of rating. Apparently handicapping doesn't do much for say, a 50 point difference in strength (you just play and take your chances). But above that, there might be ways to use handicaps. Even giving someone the first move two times out of three (as was earlier done in professional Go) might do something.

Or would it? Why hasn't handicapping been done much formally in chess, as in Go?

  • I've heard of people offering move advantages early in the game before, but I'm not sure what metric you'd use to determine how much one move is worth in rating terminology. Does a bonus 1 move for the first 3 moves equate to a 200 point spread? – James Tomasino May 1 '12 at 19:25
  • 3
    An idea would be to give the stronger player less time, but I can imagine that it is difficult to get the right ratio. – Landei May 1 '12 at 20:08
  • Landei, I remember reading a book about Tigran Petrosian that mentioned something similar in one of the tournaments he played in. Time was sacrificed not as a handicap for rating differences, but as an option to play as white when you normally wouldn't. I don't recall exactly where that was used or what the rules around it were, though. – James Tomasino May 1 '12 at 20:15
  • 3
    "... Japanese game Go" What!? – NullUserException Jan 18 '13 at 21:36
  • @NullUserException Yeah, that's a Chinese game... – ericw31415 Feb 6 at 22:45
up vote 13 down vote accepted

While giving an opponent the odds of a pawn or a knight might be fun to make the winning chances more distributed, it is by no means an exact method. I remember the exciting handicap match between Kasparov and Chapman. Kasparov won 2.5:1.5 by giving the odds of 2 pawns against an opponent of approx. 450 Elo points less. Since the 2 pawns given changed for each game, not all games were equally difficult for Kasparow. During an interview the former world champion said he feared the game most where he had to let go of the king's center pawn. So a pawn is not a pawn. And if I remember correctly he added that a two pawn disadvantage was much more than double a one pawn disadvantage.

Hence I believe you can't construct a meaningful metric system of which handicap is worth how many points.

Handicapping changes chess in a fundamental way that stones in go do not.

In go, getting an extra stone is roughly equivalent to having an extra move. In chess, losing a piece is catastrophic, and in a game between two players that are within ~300 rating points, losing a piece means loss of the game.

One other reason that handicapping is not done this way is that while you (1500 player) might lose to a 2200 player (barely) without a knight, it is highly unlikely that a 2000 player would lose to a 2700 player at a tournament time control.

Handicapping was done before rating systems became so prevalent. In clubs, players would be said to be "knight players" meaning that they needed odds of a knight in order to give a stronger opponent a good game. Similarly there were "pawn" players and "rook" players.

An easier handicap between two roughly equal opponents would be to give the weaker player white. This is roughly equal to 25-40 rating points [1]. Another common handicap is for the stronger player to play "blindfolded", meaning that he or she cannot look at the board, but has to call out the moves in order to play.

  • if the equivalent in chess would be moves, why not give a value to early game moves? A weaker player could open the game with two moves instead. There is obviously a limit to how many moves can be given up, and each move would probably have it's own value. – Justin C May 1 '12 at 19:41
  • 2
    I believe that was done before rating systems as well, although the general consensus is that a pawn is worth 3 tempi, so it was far more common to remove an f pawn from the stronger player just because the nature of the game is changed less when a single pawn is removed than when white gets two moves in a row. – Andrew May 1 '12 at 19:46
  • wonder how much handicap would be to don't allow castling, my guess is 1 pawn and giving the opponent whites (with the opponent knowing you can't castle) and 1 pawn without knowing – ajax333221 Aug 8 '12 at 19:32
  • @ajax333221 kind of random, but Deep Rybka says that if white cannot castle from the start of the game, black has an edge of 0.33 pawns. – Andrew Nov 21 '12 at 20:18

For "fun" games with players of different strengths you can use a chess clock and give the weaker player more time. This allows me to play against my son who is just learning a reasonable chance at winning. We often play where he gets 5 minutes on the clock to my 2 minutes.

  • 2
    This should be used more often, it's the only handicap that doesn't change the game. – Eric Wilson May 24 '12 at 12:20
  • 2
    A fun variation on this is to play a progressive version of this over several games (say 5). 10 minutes each to start, but the winner of each game gets a minute knocked off for the next game. – AndyM Jul 27 '12 at 11:41
  • 1
    Only it soooo much depends on the players involved, and their blitz strengths. Also, upwards from a significant difference of level, a strong blitzer can win with ridiculous amounts of time against a long-thinking opponent, and then it becomes more of an arm-fitness exercice (I won a game with 30s against 20 minutes against a 300 points weaker opponent, out of one game, but it was straightforward enough for us not to give it any more tries). – Nikana Reklawyks Nov 21 '12 at 6:56

In some ways, the rating system is the handicap. It's very similar to the odds given to less impressive horses in racing, or the less favored boxer.

If you over-perform based upon you and your opponents respective ratings, you'll have a net gain of points, even if you don't win many (any?) games.

I drew a master once and sheared him of many points. When he beat me at another match, he got like 1 or 2 points because according to the his rating it was expected.

edit, for the comment:

No, I understand the OP completely. The opponents one is paired with are based upon rating. When one's rating drops (or rises), one is paired with players in a similar skill level range. So the "handicap" is not having to play brutally better players as often. There isn't a practical way to handicap individual games in chess like there is in Shogi.

At low skill levels, players drop pieces so any sort of handicap would be pointless. Even at my cruddy level, a single pawn adds 200 or 300 rating points, and a piece would add 500 or more. There simply isn't a lot of granularity in chess handicaps.

Usually, handicaps in Go are used between a weaker and a stronger opponent to make the game more competitive, giving the weaker player more opportunities to attack and defend as would happen in an even game. This is done in turn to avoid the game from becoming a sweep by the stronger player, with the weaker player constantly fumbling defenses. The true underlying purpose of handicapping should be seen as accelerating the learning of the weaker player as opposed to making the game more enjoyable for the weaker player. If the latter is what the player wants, he should seek play with players of equal ability.

As described in Andrew's answer, unfortunately, there is no clear way to handicap a game of chess as there is in go. However, if the goal is to use handicapping as a means to accelerate learning chess between a tutor and a student, a common thing to do is to play a normal game of chess, and upon encountering a situation where the weaker player thinks he has a loosing position a switch of sides is done upon request of the weaker player: the weaker player becomes the attacker, the stronger player becomes the defender.

From personal experience, a stronger player is a better chess player not because they have some magic ability, but because they see more possibilities and understand the variations better. In this sense, the stronger player is usually able to recover the weaker player's poor position (especially given the rating discrepancy between attacker and defender) making this a very instructive and hands on learning process. As a side effect, the weaker player is given the chance to practice executing attacks correctly.

Obviously for this to work, both parties must agree that winning and loosing doesn't mean much, and it is simply play for the enjoyment of the game and to accelerate the weaker player's understanding.

For my niece and nephew (ages 5 and 7), I started spotting them four pieces of their choice ( minus K or Q) and each time they win the number decreases by one. We started by switching sides when things were going my way.

I play against kids at my high school, and they're terrible. I can handicap a queen and both rooks, but usually I still win. When you're playing against low-skill players, handicaps don't mater much.

Like some people have mentioned in the thread, both time and piece handicaps are popular in chess. You also can do "wager" handicaps if you're playing for cash or play-money currency.

There are several chess websites that now offer handicapping to help bridge the often vast skill-level differences in chess. Out of all of them I recommend:

http://www.velocitychess.com

This site allows ANY combination of piece, cash, and wager handicap. For example, as a USCF master, I may offer a weaker player an extra starting knight, his 5 minutes to my 1 minute, and 3:1 wager odds in his favor for a game. It is a lot of fun!

Besides pieces, you can also give free moves, but both change the game. Playing while down a queen isn't about how well you can play chess, it's about how well you can predict your opponent's moves. Good chess is about making moves that are good no matter what your opponent does, not about correctly guessing what blunder your opponent will make. This is similar to how in games in general, there's only so much that a handicap can make up for bad AI.

Something that changes the game less is to give one player more time, but the problem with that is that there's a limit to how much advantage that gives: at some point, the weaker is taking so much time that the stronger player is simply using the weaker player's move to think. If the stronger player can think through two half-moves in less time that it takes the weaker player to think through one half-move, then giving the weaker player more time just helps the stronger player. This effect can be reduced by having two or more weaker players play the stronger player simultaneously. Playing blindfolded can also be a handicap, but how much this hurts a player varies.

Another variant, that is similar to ldog's suggestion:

The weaker may, when making a move, call out a positive integer. If they have not done so before, it is optional, but once they call out a number, all further moves must be accompanied by a positive integer less than the previous one. If this is not possible (i.e., the previous number was 1), and the move is not checkmate, they lose.

Until the stronger player calls out a number, the weaker can switch sides/colors freely. Once the stronger calls out a number, the weaker player can switch sides at any time, but this switch is permanent, and the weaker player now has that many turns to deliver checkmate.

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.