Join Bridge Winners
All comments by Nicolas Hammond
1 2 3 4 5 6 7 8 9 ... 133 134 135 136
You are ignoring the author of this comment. Click to temporarily show the comment.
@Adam: “I think the problem is especially bad in long team matches”

I disagree.

Back to your first sentence in the OP, “Slow play is going to kill high-level bridge”.

You need to better define, “high-level Bridge”.

I think we need to separate the arguments into Pairs, Swiss, KO.

Let's take Regional events. I'll use Gatlinburg as the example, lots of events.

It is rare for a KO match to go beyond regulation time. I know because we sometimes pre-sell to the losers for the next event; you can see from the empty tables in the KO area.

I don't know much about Pairs games in Gatlinburg.

There are some slow players in Swiss events. ACBL has CoC, using proper tools (shameless plug: Bridgescoreplus), the TDs can not only enforce the time penalties, but have documentation to prove that scores were reported late. For the last few years at Gatlinburg, I've brought an atomic clock with me, just to prove the time stamps are “real time” (it also helps if someone wants to buy an entry at 1pm and TD has closed sales). I think once players know that penalties are enforced, they will improve. See previous comment about experience running Swiss with penalties and Board influence on TDs.

Let's take Nationals.

For the main events: Spingold/Vanderbilt; the earlier rounds are generally done in time. One slow pair will only slow up that match. This should be a separate topic from other events.

Reisinger: The first day is not Barometer. Can't remember about second day. Third day is Barometer. I did Vugraph for the last Reisinger. There were some slow players. I don't know if any warnings or not; but I do know that some players were complaining about missing their breaks. With the last day of the Reisinger, you are down to small group of players that are general self-policing. Rather than an entire BW discussion, including many who never will play on the final day, this is probably best done by discussing with those that play at that level.

For the NABC+ pair events: this is where I think the major problem exists. We have a limited amount of time per round; if you are following a slow pair, it is very frustrating. Particularly as you will probably be following them in the second session as well. I've got to the point where I write down the number of minutes left in the round on my scorecard when I follow a slow pair; inevitably there is one round where we are “slow” (we started late), and the TD wants to issue a warning, but nothing is ever done to a “top” pair that is slow.

Fast pairs are generally handled well. All players and TDs know that penalties are issued. One fairness issue would be to have cameras on all tables and players would have the ability to appeal a late play if it can be shown that they were not responsible (slow uncontested auction by opponents taking 5 minutes to 6H).

Bottom line: I'd like to see TDs enforcing the rules. I'd like to see ACBL have CoC for NABC+ pairs events that document the penalties (Fast Pairs has this). TDs have the tools (BWS data).
Feb. 20
You are ignoring the author of this comment. Click to temporarily show the comment.
The ACBL has regulating authority. The Laws don't prescribe a penalty. The ACBL CoC for Swiss events does. The ACBL Fast Pairs CoC prescribes the penalties.
Feb. 19
You are ignoring the author of this comment. Click to temporarily show the comment.
There are different merits to each system; but an issue is which is better to teach to new players.
Feb. 19
You are ignoring the author of this comment. Click to temporarily show the comment.
@Peter: I made the suggestion (to the right person) during the 2018 Orlando WBF events to change the order for subsequent rounds. It's a simple coding issue. Either randomize, or sort in reverse order (fixes for one round). It wasn't going to be fixed at Orlando, but should be for subsequent WBF/EBL (same software) events. ACBL is a different story.
Feb. 18
You are ignoring the author of this comment. Click to temporarily show the comment.
Hardy is definitive 2/1; but he includes many treatments that almost no-one uses. Lawrence has a derivative 2/1. Main difference between the two is “strength” v. “length”. In a 2/1 auction, does opener's rebid show strength or length. Example: 1-2-2. Does 2 promise 6 spades or not?
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
The ACBL has a book that beginners can buy. It is written by Audrey Grant. I give it out to my class. It has 4 chapters. First chapter is general; second is 1NT openings, third is major suit openings, fourth is minor suit openings. This would be “beginner bridge”. No transfers.

SAYC is what you are supposed to have to play if you do not have two completed convention cards at the table. See http://web2.acbl.org/documentlibrary/play/sayc_card.pdf

A “pickup partnership default” would assume that all teachers that teach beginners give the same class. This doesn't currently exist in ACBLland.

You could argue that there is some benefit to a “beginners card”, “sayc”, “2/1”, “expert 2/1” that is prefilled out and players could make changes to when filling out a card.

Also, players learn different conventions at different times. Trying to standardize on the conventions that are taught is also difficult. I would argue that Blackwood, Stayman, Transfers are the minimum. Others disagree.

@Chris: I set homework. And expect them to do it. I'm the mean teacher.
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
For egregious late plays the TDs will sometimes put a “dot” over the team name on the Jeffrey's Chart to indicate the number of late plays issued.
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
Click on
http://www.bridgebase.com/tools/handviewer.html?d=n&v=0&w=skt763h96djt74c98&n=s85h753dak9632ck2&e=sj4hkt42dq85cj753&s=saq92haqj8caqt64&a=PP5C(Generated)PPP
to see this hand on BBO. You can then use GIB to find the solution.

Nice hand. Order of cashing tricks is NOT obvious.
Feb. 17
Nicolas Hammond edited this comment Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
The ACBL really has nothing to do with “pickup partnership default”.

Most Bridge teachers start Beginners with SAYC, no transfers.

I just finished a 6 week beginner class. I taught 2/1, Blackwood, Stayman, 2 level major suit transfers, std carding/discards. Other teachers at the same club teach SAYC, no transfers.

There is no “right” or “wrong”.

Most beginners start out with SAYC and “graduate” to 2/1.
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
“@Nick - bridgemates solution is unrealistic. Haven't you ever forgotten to score a board till you were into the next one? If you are going to have a real slow play enforcement policy you need something that is totally objective and totally non-discriminatory.” (FYI: I prefer Nicolas not Nick).

Bridgemates were considered unrealistic for pairs and team games. Until players started to use them. On the rare instances (they do happen, but they are rare), when a player forgets to confirm a score (thought they pressed the confirm button, but didn't), or when the TD/DIC/ACBL is responsible (think round 1 when the Bridgemates are not active until almost the end of round 1), give the TD the ability to waive. If a pair gets a late penalty. These should be the rare exception.

What can ACBL do?

1. For pairs events publish what the late penalties are.

The CoC are here: https://www.acbl.org/tournaments_page/charts-rules-and-regulations/conditions-of-contest/

The Swiss CoC, http://web2.acbl.org/coc/SwissGeneral.pdf (this appears to be a more recent version than the one I referred to earlier), does list the penalty. Warning for first offence, 10% of highest possible match score for second offense, 20% for third, 40% for fourth etc.

There is currently no prescribed penalties for late plays in pairs events.

2. Enforce.

For pair events, the data is in the Bridgemates.

For Swiss events, use Bridgescore+. The time that each round is started is known (the time of the last assignment being posted), there is a running clock on the projector, the time that the match is reported is recorded. Bridgescore+ runs the events faster than ACBLscore, and needs fewer TDs. Bridgescore+ is certified by ACBL. It has run events at NABCs.

3. Change the session 2 matching algorithm (see upthread).

Very frustrating being behind a slow pair in session 1 to know that you will have the same problem in session 2.

I've followed very slow (and famous) pairs at NABC+ pairs events. One time I complained 10 times in 13 rounds because we had to wait. Including after one break round! Nothing ever said to the other pair. No penalties ever issued.

Edit: Added more in (1)
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
To be completely correct, it may have been an MABC board member, not a D7 board member. The distinction is lost for those not from D6/D7. (Quick summary: MABC is the entity that used to run the tournaments in D6/D7). The gentleman (for it was a he, not a she) has, I think, been on both boards. At the time I don't know which board he was on. I've been on the D7 board; I seem to recall he was also on the board when I was on it. But when this happened he may have been on the MABC board, or the D7 board, or both.

D7 can always repeat the experiment. Run Swiss with Bridgescore+ and enforce the late plays. D7 keeps track of the time that all scores are entered. The clock is supposed to start when the last assignment of the previous round is posted. ACBL gives you time to shuffle/deal/play/score/report. If the report is not presented when the clock runs out, you are late. See page 3. https://cdn.acbl.org/wp-content/uploads/2014/02/Swiss-Teams.pdf. Despite the numbers in the URL, this is the 6/20/2016 version. I don't know if this is the latest.
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
Not mine to name. I was helping the TD run the event with Bridgescore+. I am not an ACBL TD or TA. Because I am often close to the TD station, I cannot help but overhear some conversations, or the TD will share information with me about the event and the reasons decisions are made. The ultimate running of the event is up to the TD/DIC. What actions the TD/DIC took afterwards are things I am not privy to, nor should be privy to.
Feb. 17
You are ignoring the author of this comment. Click to temporarily show the comment.
District 7 implemented this in a Swiss event. Two session. Plenty of warnings given out before the event started and during round 1. Event was run with Bridgemates and Bridgescore+.

Automatic assignment of penalties following the ACBL rules.

There was one team that was late in rounds 1, 2 and 3. A District board member was on the team. He has self-convinced that he is a fast player. Yet, he is a slow player. Every round he blamed the opponents for being slow. Yet each of the teams he played against only had one penalty (when they played his team). He complained so much at Round 4 and during the break that the DIC removed all penalties for the second session.

The District has not repeated this trial.

All the other players loved it; except the one District board member. He was powerful enough to kill the implementation.
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
@Robb: “Regulating authorities have tried just about everything except waterboarding (covert observation and penalty, warning and penalty, begging, prohibiting slow pairs from playing together on teams etc).”

Actually ACBL have tried very little. Bridgemates record the time stamp of each board. Trivial to auto-assign penalties. You could publish video and if a pair can show they were not responsible, their penalty is removed. Notify pairs that have been auto-assigned a penalty by having a director drop off a pre-printed notification during their next round.

Both ACBL and WBF fail in second session assignments. If you follow a slow pair in the first session, you are likely to follow them in the second session. This is a known problem, reported many times, that remains unfixed. WBF said they might fix it. Fix it to keep the section assignments for session 2 but reverse the order of the pairs.

Neither of these simple fixes have been tried or implemented.

And if you want software that does the automatic assignment of penalties, let me know :-)
Feb. 15
You are ignoring the author of this comment. Click to temporarily show the comment.
You are ignoring the author of this comment. Click to temporarily show the comment.
Perhaps the better people to ask are those that have won a National event.
Feb. 12
You are ignoring the author of this comment. Click to temporarily show the comment.
As an example; I took the top 200 pairs (with data that excludes all known cheating pairs) based on amount of data from top level play. I have the most data on Meckwell. Their difference is 0.9%, the other Geoff/Eric is also 0.9%. Bobby/Steve 0.7%. Katz/Nickell 0.9%, Multon/Zimmerman 0.8%, Rosenberg/Zia 0.8%.
Feb. 10
You are ignoring the author of this comment. Click to temporarily show the comment.
I have all the statistical evidence. I am afraid you have a mis-perception.

After the opening lead is faced, the defense has an advantage over declarer, of approximately 0.7%; assuming that all players are of a similar ability.

Your perception may be different - if you are a strong player, and play against weaker players, you will notice more of the opponents errors.

For any given player/pair, then will make 0.7% more “mistakes” when declaring than they will on defense after the opening lead.

Each of the defenders can also see 26 of the 52 cards. Also, typically, the declarer has shown a HCP range and some shape giving away more information about their hand.
Feb. 10
You are ignoring the author of this comment. Click to temporarily show the comment.
Ed: From the Laws themselves:

Established usage has been retained in regard to “may” do (failure to do it is not wrong), “does”(establishes procedure without suggesting that violation be penalised) “should” do (failure to do it is an infraction jeopardising the infractor’s rights but not often penalised),”shall” do (a violation will incur a penalty more often than not) “must” do (the strongest word, a serious matter indeed). Again “must not” is the strongest prohibition, “shall not” is strong but “may not” is stronger – just short of “must not”.

You can have a valid game of Bridge if the backs are not identical. However, if the wording was “must be identical”, you cannot have a valid game of Bridge according to the Laws.
Feb. 10
You are ignoring the author of this comment. Click to temporarily show the comment.
@Tomas: “Guys, a few years ago a Polish player has made a crude statistic. He took only boards without defenders bidding and checked how often the first lead found partner's suit with one of A/K/Q.

Among pairs with enough boards in top international events first three places were:
Fisher-Schwartz, Elinescu-Vladov (and these two pairs had HUGE advantage), Piekarek-Smirnov.

B-Z were in the middle of the pack.

I know who had the fourth place but decorum forbids me etc. :)
All I could say that they played 85% of boards on NS line (Fisher-Schwartz only 65%).”

Tomas's comment is up thread. I've moved my reply to the current end of the discussion as the thread it was on was getting crowded.

I have the same statistic. Back in 2015, there was both a Polish spreadsheet, and also, I think, one from India. Unfortunately in both cases it was impossible to get the original data used for the spreadsheet, so it was impossible to verify the information and thus impossible to use for any of the hearings. I write about opening leads, including this test, in my book: p109-128. In one claim from back then, F/S were alleged to have led to partner's honor in an uncontested auction 86% of the time, while other pairs were in the 55% range. The raw data was never forthcoming and so this value could not be checked or used. The data I had (which I shared with EBL) contradicted this alleged number. As I shared all my raw material, and the results could be verified, EBL did not use the other data.

A difficulty with this work is deciding which tournaments to use, and then what the minimum threshold for number of leads to be included. My data is “top” tournaments - I have about 300 in my database. I don't use player provided opening leads; Vugraph only.

Another difficulty is: does this show cheating. Is a “passive” leader, one who never leads away from an honor likely to show up on this list because partner is more likely to have an honor in the suit?

In my book, I use top tournaments, top 100 players based on amount of data, non-competitive auctions, non-trump lead, pre summer 2015 tournaments only. As this is statistical information, it is publishable. Finding partner with A, K or Q in the led suit:

1. Fisher/Schwartz. 63.1%
2. Buratti/Lanzarotti 62.2%
3. Wladow/Elinescu 61.1%
4. Smirnov/Piekarek 60.7%
5. European pair 58.8%
..
7. American pair: 57.7%
8. Fantoni/Nunes 56.35%

The other places in the top 10 were European pairs. In my book, I do not name the players (except those listed above). In most cases I list the continent they are from.

There is a different philosophy on leads between European and American players. Europeans generally make better leads than Americans.

I asked some top experts what they would consider on opening leads. If they had a lead from AK, KQ, QJ10 or QJ9 in an uncontested auction they would lead that suit.

Another way of doing this statistic is to only include boards where partner has a non-trump ace, king or queen. It changes the results slightly. (i.e. exclude boards where the partner of the opening lead does not have any A, K or Q outside of the trump suit).

As an example: this table is all auctions (including competitive auctions), non-trump leads, partner has an A or K (ignores the Q). Opener does not have one of the suit combinations above:

1. B/L 59.9%
2. F/S 59.6^
3. European pair: 59.5%
4. P/S 59.4
5. Wladow/Elinescu: 57.6%

F/N were #28 at 53.2%, BZ were #31 at 53.1%. Stewart/Woolsey were at #39 at 52.3%. I list Stewart/Woolsey as they were allowed me to use their names in my book. They are a good reference pair.

Supposing I increase the data set to the top 200 pairs. Now we have
1: 63.6%
2: 62.3%
3: 60.2%
4 B/L
5: F/S
6: 59.4%
7. PS
8: American pair - 59.3%
12: W/E

Does this mean #1, #2, #3 were cheating; or, is this the general law of statistics where you need a large number of boards. This is where the true mathematicians and statisticians get involved and Bridge players need to step aside.

My book lists various different ways of calculating similar information, but excluding certain leads.

One of the most interesting results was non-competitive auctions, non-trump leads, all boards irrespective of opener's holding, finding partner with an A, K, or Q. This is the same table as the original; only I included data from more US tournaments, so there were more US players in the top 100:

1. <American pair>.: 63.8%
2. F/S 63.1%

14. Hamman/Soloway: 56.1%

Does this mean that the #1 ranked American pair was cheating? All I can state is that things changed for that pair after 2015. But little/no video on them pre-2015, so put it down to climate change.

If I increase the number of pairs, based on data, from 100 to 200, then… F/S drops to #5 on the list. Going from 100 to 200 pairs starts to include some of the pairs referred to in the OP. (Not going to say if any of those pairs are in the top 4 spots!). The #1 on this new list has a clear lead over #2. Alas, little/no video on this pair.

If you were to ask a Bridge player, absent knowledge of the data above, what holdings they would like to signal to partner for an opening lead, assuming such signals were legal, they would include hands with Ace or King; but they don't include hands that are just the Queen. For example, Qxxx is not a suit that you want partner to lead to. At least, that was the perception when I talked to top players. Therefore… should you be using the statistic of A, K or Q; or just A, K?

The real advantage to being able to signal for an opening lead, is not necessarily that this is the suit you want led; it is the fact that partner has signalled strength in that suit. That information can be used later in the hand.

Opening lead statistics are difficult to use for detecting cheating. The simple percentage of times that partner has a A, K or Q in non-competitive auctions lists some of the known cheating pairs. But lists other pairs as well. If you believe that this statistic implies cheating on the opening lead, then there is at least one European pair and one American pair that you should be looking at. Only little/no video on those pairs from before 2015. If this pair had a sufficient number of boards to be statistically viable is outside the scope of this discussion.

One advantage to the OL statistics, is that it is possible to rule out a pair as cheating on the opening lead. Or… put a better way, what they are currently doing does not indicate cheating.

And… amazingly … the opening lead strategy has changed for a small number of pairs since 2015. One of the more interesting statistics is comparisons of top pairs from before 2015 and after 2015. There were some serious opening lead discussions around September 2015 for some pairs.

IMO, the AK or Q test is not the best way to detect cheating on the opening lead. I have better formulae, but these are only available to licensees of the technology.

I just wanted to set the record straight about the statistics on the AKQ test for pre-2015 tournaments. Although it identifies some of the known cheating pairs, it also strongly implies that some of the known cheating pairs were not cheating on the opening lead. Which… is a little surprising. If you decided to collusively cheat, the most obvious place to want to cheat is on the opening lead. Perhaps those who collusively cheated, but did not cheat on the opening lead, can provide more information.

Big question: are some players cheating now on opening lead? Hmm… there are some pairs that are statistically better than F/S in the AKQ test, just looking at data from post summer 2015 tournaments, however, there are a limited number of boards on these pairs therefore, until these pairs have more data, no conclusions can be drawn. These type of statistics require a large number of boards to be valid. Exactly how many is open to discussion. For me, not enough boards on them.
Feb. 10
1 2 3 4 5 6 7 8 9 ... 133 134 135 136
.

Bottom Home Top