You are ignoring the author of this comment. Click to temporarily show the comment.

And in many instances club members/owners can pay an extra dollar, hold a charity game, and dole out a multiple of the masterpoints they would have otherwise received.

You are ignoring the author of this comment. Click to temporarily show the comment.

I would think per hand vs per round would not make a great deal of difference based on the PR calculation. Per board would be a bit more accurate in case you didn't play all the opponents. If you played all the opponents it would make zero difference. And over the course of enough boards, you would expect to miss an average quality opponent so that factor should be negligible.

You are ignoring the author of this comment. Click to temporarily show the comment.

He must get information that ends up on ACBL Live - tournament information. But not club level information. I'm trying to send summaries every month or two, it's quite easy. That's why you'll find a lot of information for Des Moines (216) and almost nothing for Omaha (241)

You are ignoring the author of this comment. Click to temporarily show the comment.

All this worry about variance is overblown. Sure, you can be unlucky for a game or two, but over the course of 2 years the power ratings is going to accumulate enough good luck and bad luck that it will offset and get you pretty close to your true performance.

You might have a standard deviation of 5% over a single game, but if you average 100 games together, your standard deviation is now 0.5%. 200 games? 0.35%.

You are ignoring the author of this comment. Click to temporarily show the comment.

>Surely, a system that weights performance in NABC open Pairs the >same as a 4 table open game with no life masters cannot be an >accurate reflection of skill, can it?

It uses degree of difficulty (DoD) to put these events on a level playing field. Suppose I have a rating of 60:

If I enter the Silodor Pairs, the average opponent probably has a rating of 60 (DoD +10.0). Therefore all I need to do is score 50% to maintain my rating. 50.0% + 10.0% = 60.0%

If I play against NLMs, the average opponent might have a rating of 45 (DoD -5.0). Now i need to score a 65% in order to maintain my rating. 65.0% - 5.0% = 60.0%

As for the 4-table piece of the equation, it doesn't change the expected value of the score, just the variability. You are expected to score 65% regardless of whether there are 4 tables or 13.

You are ignoring the author of this comment. Click to temporarily show the comment.

Suppose I play two 2-session regional pairs events against identical opposition.

I get a 60% and win for 25pts, and the next day I get a 50% and earn nothing. I got 50% of the available points.

My friend gets a 55% each time and wins 5pts each time, getting 20% of the available points.

My other friend gets a 50% and wins nothing. The next day, they score a 60% but it is only good enough for 2nd overall so they get 18.75pts. So they finish with 37.5% of the available points.

These pairs all achieved a 55% average against identical opponents yet their results are wildly different.

You are ignoring the author of this comment. Click to temporarily show the comment.

I think it would not be difficult to rate IMPs as well using an additive system that measures IMPs won per board. If I win 1 IMP per board against your average player, my rating would be +1.00; if you wanted to convert that to a sexy rating number let's assume that would be 60.0

So my team of all 60's plays a team of all 50's. We would be expected to win 1 IMP per board. If we tie the match, we have both played like 55's. If we win an 8-board match by 8 IMPs (1 IMP/board), then we played like a 60 and they played like a 50. From there you can extrapolate; a 2 IMP/board win would mean we played like a 65 and they played like a 45.

From there you have to give each of the three matches in the session 1/3 weight to come up with the performance rating for that session.

In Swiss team events it is normal to receive these scores. In KO matches however, all that gets reported is a win/loss. I would probably exclude this information because there isn't much incentive for it to be reported properly.

You are ignoring the author of this comment. Click to temporarily show the comment.

The power ratings site uses two years of data which prevents a single down day from having a noticeable impact. If you play once a week then you have 104 results over the course of two years, and every month 4-5 would drop off and 4-5 would add on. It would only A new month of scores would only get 4%-5% weight. A single game would get 1% weight. If you were a 60.00 and you had a bad day and only performed at a 50.00, then your new rating is 59.90. Not exactly the end of the world.

You are ignoring the author of this comment. Click to temporarily show the comment.

Yes, dealing with teams would be more problematic. I think your proposal has some major gaps though. First and foremost, people aren't going to get to clubs, sectionals, regionals, and nationals. Assigning zeros for having an insufficient amount of data is not going to work.

>Chris tends to rank by how much a player is overrated, allowing >for an unforeseen denominator effect.

I think this is just some side information but not part of the actual calculation. If you have a 60.0 rating then he thinks you play like someone with X masterpoints. If your masterpoint total is only Y, then X / Y is his measure of how overrated you are. But, it doesn't factor in the actual rating calculation.

More importantly, the strength of field can be accurately calculated. The average opponent at NAP this weekend was 60.9. I am a 61.92 and I partnered with a 63.27 so our average was 62.60. As long as we could average 62.6 - 10.9 = 51.7% in this field we would retain our ratings. As it is, we had a 52.2% so our performance was a 52.2 + 10.9 = 63.1, so those four games will say that our rating should be 63.1/62.6 = 0.8% higher.

You are ignoring the author of this comment. Click to temporarily show the comment.

Chris Champion who authored one of those letters to the editor would seem to have the matchpoint rating system already up and running. All he needs is more clubs to report their results. Does anyone have an issue with it?

You are ignoring the author of this comment. Click to temporarily show the comment.

X and rebid 1NT is 19-20, quite an overstatement, though I don't know if you adjust it for the fact that partner didn't bid.

It isn't always easy converting diamond length into tricks in NT. Give partner

♠ xx ♥ Jxx ♦ AJ98x ♣ Kxx

I think my double is showing extras and probably a willingness to defend. Partner can figure out something logical if he doesn't want to defend ie. 1NT. If it is still extras with takeout for the black suits that works as well.

Rich Newell

Rich Newell

Rich Newell

Rich Newell

Rich Newell

Rich Newell

You might have a standard deviation of 5% over a single game, but if you average 100 games together, your standard deviation is now 0.5%. 200 games? 0.35%.

Rich Newell

>same as a 4 table open game with no life masters cannot be an

>accurate reflection of skill, can it?

It uses degree of difficulty (DoD) to put these events on a level playing field. Suppose I have a rating of 60:

If I enter the Silodor Pairs, the average opponent probably has a rating of 60 (DoD +10.0). Therefore all I need to do is score 50% to maintain my rating. 50.0% + 10.0% = 60.0%

If I play against NLMs, the average opponent might have a rating of 45 (DoD -5.0). Now i need to score a 65% in order to maintain my rating. 65.0% - 5.0% = 60.0%

As for the 4-table piece of the equation, it doesn't change the expected value of the score, just the variability. You are expected to score 65% regardless of whether there are 4 tables or 13.

Rich Newell

I get a 60% and win for 25pts, and the next day I get a 50% and earn nothing. I got 50% of the available points.

My friend gets a 55% each time and wins 5pts each time, getting 20% of the available points.

My other friend gets a 50% and wins nothing. The next day, they score a 60% but it is only good enough for 2nd overall so they get 18.75pts. So they finish with 37.5% of the available points.

These pairs all achieved a 55% average against identical opponents yet their results are wildly different.

Rich Newell

So my team of all 60's plays a team of all 50's. We would be expected to win 1 IMP per board.

If we tie the match, we have both played like 55's.

If we win an 8-board match by 8 IMPs (1 IMP/board), then we played like a 60 and they played like a 50.

From there you can extrapolate; a 2 IMP/board win would mean we played like a 65 and they played like a 45.

From there you have to give each of the three matches in the session 1/3 weight to come up with the performance rating for that session.

In Swiss team events it is normal to receive these scores. In KO matches however, all that gets reported is a win/loss. I would probably exclude this information because there isn't much incentive for it to be reported properly.

Rich Newell

Rich Newell

>Chris tends to rank by how much a player is overrated, allowing >for an unforeseen denominator effect.

I think this is just some side information but not part of the actual calculation. If you have a 60.0 rating then he thinks you play like someone with X masterpoints. If your masterpoint total is only Y, then X / Y is his measure of how overrated you are. But, it doesn't factor in the actual rating calculation.

More importantly, the strength of field can be accurately calculated. The average opponent at NAP this weekend was 60.9. I am a 61.92 and I partnered with a 63.27 so our average was 62.60. As long as we could average 62.6 - 10.9 = 51.7% in this field we would retain our ratings. As it is, we had a 52.2% so our performance was a 52.2 + 10.9 = 63.1, so those four games will say that our rating should be 63.1/62.6 = 0.8% higher.

Seems logical to me.

Rich Newell

bridgepowerratings.com

Rich Newell

Rich Newell

I upgrade lots of 17's (minor or major opening) with decent 5 card suits and controls.

But not

♠ AQ5 ♥ K5432 ♦ A6 ♣ KJ2

Rich Newell

Rich Newell

Rich Newell

Rich Newell

It isn't always easy converting diamond length into tricks in NT. Give partner

♠ xx ♥ Jxx ♦ AJ98x ♣ Kxx

I think my double is showing extras and probably a willingness to defend. Partner can figure out something logical if he doesn't want to defend ie. 1NT. If it is still extras with takeout for the black suits that works as well.

Rich Newell

Rich Newell