[riverwalk3]

Imagine a scoring system where one got +1 point for each level that they bid, assuming the contract is made (making their score not a multiple of 10 in that case). Thus, if NS bid 4H nonvulnerable and made exactly, they get 424 points. If they make 5 instead, they get 454 points. If they bid 5H and make 5, they get 455 points.

I wonder how matchpoints would look like with this scoring system.

[mikeh]

 riverwalk3, on 2022-May-26, 21:36, said:

Imagine a scoring system where one got +1 point for each level that they bid, assuming the contract is made (making their score not a multiple of 10 in that case). Thus, if NS bid 4H nonvulnerable and made exactly, they get 424 points. If they make 5 instead, they get 454 points. If they bid 5H and make 5, they get 455 points.

I wonder how matchpoints would look like with this scoring system.

Weird


And it wouldn’t work as you may be thinking it might.

Say 4H is cold but 5H is ‘only’ 75%. Nobody in their right mind would contract for 5H. Unless slam was a realistic possibility, and often one knows it isn’t, reaching 5H would likely be the result of bad bidding. Yet if it makes, they beat the better players who played 4H. That’s too random and distorting.

[sfi]

 riverwalk3, on 2022-May-26, 21:36, said:

Imagine a scoring system where one got +1 point for each level that they bid, assuming the contract is made (making their score not a multiple of 10 in that case). Thus, if NS bid 4H nonvulnerable and made exactly, they get 424 points. If they make 5 instead, they get 454 points. If they bid 5H and make 5, they get 455 points.

I wonder how matchpoints would look like with this scoring system.

I've long wanted to try out a related idea - you get 5 bonus points for making the contract exactly. Your idea is a bit more precise than mine and it would certainly be an interesting change.

[smerriman]

 mikeh, on 2022-May-26, 22:34, said:

Nobody in their right mind would contract for 5H..
..reaching 5H would likely be the result of bad bidding
.. Yet if it makes, they beat the better players who played 4H.

Well, those statements are only true if you use current bidding systems, which make no sense under this scoring. Changing the scoring entirely means you'd need to develop a completely different bidding system to be a "better player.".

But that seems far too much work.

[LBengtsson]

There's an American expression that I learnt from my mechanic: If it is not broken, do not try to fix it. Bridge scoring has been around for many years in its present form. It will just add a complication to something that does not need more complication.

[AL78]

 mikeh, on 2022-May-26, 22:34, said:

Weird


And it wouldn’t work as you may be thinking it might.

Say 4H is cold but 5H is ‘only’ 75%. Nobody in their right mind would contract for 5H. Unless slam was a realistic possibility, and often one knows it isn’t, reaching 5H would likely be the result of bad bidding. Yet if it makes, they beat the better players who played 4H. That’s too random and distorting.

I'm inclined to agree. Last week I played against a pair where one opened 1NT, the other had 5-4 in the majors, they went through Stayman and ended up in 3NT, completely missing their 5-3 spade fit, and 3NT makes the same number of tricks as 4S so they get a near top. More randomness is not needed in the game.

[pilowsky]

The unusual thing about Bridge scoring is that it is unrelated to the optimal result of the game.
In most other equilibrium games where each player/pair starts from a set position there is an optimal result depending on the scoring method.
In chess it's simple: you score 1 if you checkmate your opponent or make their life so difficult that they use up all their time.

Bridge is an untimed game, for no obvious reason, so it is possible to end up with an adjudicated result - often to the detriment of the pair that was ahead.

Unlike most other games that are similar to Bridge (go, chess, chequers, backgammon etc etc) there is an optimal result that can be known from the start (if you knew where all the cards are).
Given that this is the case, and setting aside the untimed problem, I believe that an optimal scoring system in Bridge is one that rewards the pair that bids (and plays obviously) as close as possible to equilibrium point of the hand.

Bridge in its current form is a very new game.
The 'not broke don't fix it' argument is not applicable.
The fossil-fuel powered motor car has been around slightly longer; it's definitely broken and it's breaking everything else around it.

[AL78]

 pilowsky, on 2022-May-27, 03:42, said:

The unusual thing about Bridge scoring is that it is unrelated to the optimal result of the game.
In most other equilibrium games where each player/pair starts from a set position there is an optimal result depending on the scoring method.
In chess it's simple: you score 1 if you checkmate your opponent or make their life so difficult that they use up all their time.

Bridge is an untimed game, for no obvious reason, so it is possible to end up with an adjudicated result - often to the detriment of the pair that was ahead.

Unlike most other games that are similar to Bridge (go, chess, chequers, backgammon etc etc) there is an optimal result that can be known from the start (if you knew where all the cards are).
Given that this is the case, and setting aside the untimed problem, I believe that an optimal scoring system in Bridge is one that rewards the pair that bids (and plays obviously) as close as possible to equilibrium point of the hand.

Bridge in its current form is a very new game.
The 'not broke don't fix it' argument is not applicable.
The fossil-fuel powered motor car has been around slightly longer; it's definitely broken and it's breaking everything else around it.

Bridge is timed in that you are expected to play a board in around 7 minutes, but the unusual thing about the game is that you often have no influence over your opponents. This means you can bid and play perfectly on a hand and get a poor result. The optimal scoring system in bridge would be for each board, to take each pair, look at just the factors they could control, and based on those factors, how did their score compare to the best they could possibly do. This would reward pairs for good bidding and play and eliminate being fixed by the field or the opposing pair (i.e. it rewards skill, penalises mistakes, and doesn't reward or penalise for luck), but would be very difficult to implement in practice.

[pilowsky]

Here's a common example of the problem in tournaments with a spread of ability.
This was generated randomly.

In the average field hands like this generate all kinds of outcomes.

---

The "winner" will most likely be the NS pair in 4S=. Other tables in the optimum contract of 4HX-1 will "lose".
Results like this suggest that the scoring system fails to truly reflect the best outcome of the game.
This unfair element of luck based on what other people do suggests that the current tournament scoring structure is fundamentally flawed.

I don't know what the solution is but a tournament structure that rewards luck instead of judgement seems wrong.

[riverwalk3]

 pilowsky, on 2022-May-27, 06:33, said:

Here's a common example of the problem in tournaments with a spread of ability.
This was generated randomly.

In the average field hands like this generate all kinds of outcomes.

---

The "winner" will most likely be the NS pair in 4S=. Other tables in the optimum contract of 4HX-1 will "lose".
Results like this suggest that the scoring system fails to truly reflect the best outcome of the game.
This unfair element of luck based on what other people do suggests that the current tournament scoring structure is fundamentally flawed.

I don't know what the solution is but a tournament structure that rewards luck instead of judgement seems wrong.

That probably is especially bad in express tournaments, when you have to deal with random partners and opponents. Oftentimes you don't have that much control over the outcome, as your opponents matter much more. Robot tournaments are probably the least bad in this regard.

But even in Robot tournaments (where variance of the opponents are eliminated), luck will always be rewarded in the short run. If you take a line that's 50% to make and an expert takes a line that's 65% to make, but there is a chance that you make and the expert doesn't and that chance happens, then you get rewarded.

The scoring idea was to increase the fun of matchpoints. Matchpoints deviates from the spirit of the game anyway (where every little overtrick counts a lot), so I might as increase the fun by making people bid the contract exactly. If you want sound bridge, play total points or IMPs.

[Douglas43]

This is basically a development of the main difference between Auction and Contract -that you are credited "below the line" only for what you bid. The downside is that it would penalise (say) the opponents who bid 5♣ and finish up defending 5♥= instead of 4♥+1. So I think we might see more one-sided auctions as pairs balance less and compete less.

[wuudturner]

It feels like this is an attempt to fix something that is not broken. The result will be a subtle bias in how we would bid and compete. In the end, it would change the game but I'm not sure why that change would be a positive one.

Suppose I stop in 2♡, making +1. My score will be 142.

At table 2, they get to 3♡, making the same 9 tricks. Their score will be 143. This rewards the pair who pushed to a more dangerous level, even if that extra bid was a bad idea.

On the other side, Douglas43 points out that if you compete against your opponents and push them up to a higher level, then you can be punished for forcing them to bid once more.

So, yes, it would make the game different, but where is the gain?

[riverwalk3]

 wuudturner, on 2022-May-30, 17:48, said:

It feels like this is an attempt to fix something that is not broken. The result will be a subtle bias in how we would bid and compete. In the end, it would change the game but I'm not sure why that change would be a positive one.

Suppose I stop in 2♡, making +1. My score will be 142.

At table 2, they get to 3♡, making the same 9 tricks. Their score will be 143. This rewards the pair who pushed to a more dangerous level, even if that extra bid was a bad idea.

On the other side, Douglas43 points out that if you compete against your opponents and push them up to a higher level, then you can be punished for forcing them to bid once more.

So, yes, it would make the game different, but where is the gain?

The point is that it's supposed to fit the spirit of matchpoints. You are supposed to risk going down in game to make an overtrick if you have a 60% chance of success, for example. Likewise, if 5S has a 60% chance of making in an uncompetitive auction, you should bid 5S instead of 4S.

IMPs would remain unaffected as a 1 point difference is nullified.

[TylerE]

 riverwalk3, on 2022-May-30, 18:25, said:

Likewise, if 5S has a 60% chance of making in an uncompetitive auction, you should bid 5S instead of 4S.

In a vacuum, maybe. But in real life, on your hypothetical board, there will be several pairs either missing game or in slam going down, so the odds of 5M are really much worse than you're portraying.