Explanation of ACBLmerge Report Features

See explanations below for each numbered item on the picture.

Markup of ACBLmerge reporting features


ACBLmerge is a free (open source) program developed by Matthew Kidd to merge hand records, double dummy results, and electronic scoring data with ACBLscore output, creating HTML output for websites. ACBLmerge was inspired by the program Evan Bailey wrote to merge hand records with the results generated by ACBLscore, the DOS based program used by directors. His software was used at Adventures in Bridge until 2012 when the club switched to ACBLmerge. The ACBLmerge program provides greatly enhanced functionality.

Club or tournaments desiring an easy way to use ACBLmerge with full calendaring functionality and social networking features should consider the Bridge Results web site. Read more about Bridge Results.

The numbered sections below refer to the sample output shown above.

To obtain the ACBLmerge software click here.

Send feedback and bug reports to software@lajollabridge.com

1. Top Header

The top header contains several hyperlinks. Both the green question mark and the first link both take you to the page you are reading. The PBN links allows you to save all the boards in Portable Bridge Notation (PBN) format. Many programs can read PBN format, for example the free Bridgify double dummy solver. Having the boards available in PBN format saves a lot of typing. The third link allows you to save all the boards in Goren in a Box (GIB) format. In some cases, a link to the a Dealmaster Pro generated PDF file may also be present.

You may jump directly to any board via the in-page navigation links. After jumping directly to a board, you can copy the URL from the browser and include it in e-mail or any other document which supports hyperlinks. For example, the following URL with the #board_results4 at the end will take you directly to board 4 of the September 14, 2008 Unit Game.

2. Field Strength Calculation

If field strength calculations have been included, the average (mean) and geometric mean of the masterpoint holdings of all ACBL players in the field will be shown. In a geometric mean, the mean is taken in a logarithmic manner. For example, the geometric mean of two players with 10 and 1000 MP respectively is 100 MP rather than 505 MP. Less experienced players drag down the geometric mean more than the arithmetic mean. A field with a high geometric mean should in principle be uniformly tough and offer few gifts. I have argued that geometric mean is probably more meaningful, at least for the distribution of players at tournaments. Note: in computing the geometric mean, all ACBL players are assumed to have at least 1 MP; this avoids taking log(0) and the downward skewing caused by players with only a small fraction of a masterpoint.

The field strength bar provides a visual indication of the field strength. The axis is logarithmic from 1 MP - 20000 MP. A 0-5 newcomer game would appear on the far left side of the bar and an NABC+ event would be on the far right side of the bar.

3. Masterpoint Winners

If a list of masterpoint winners has been included, all partnerships winning masterpoints will be shown, ordered by descending percentage. Codes such as OA or SB indicate the basis for the award; for example from placing in the overalls (O) of Flt A or the section (S) of Flt B. Players always receive the highest award they qualify for without vacating their ranking in the other flights they are qualified to be ranked in. Example: if the award for 1st overall in Flt B is greater than the award for 5th overall A, the pair will be shown as 1st (OB) and no pair will be shown as ranking 5th overall in Flt A.

The percentage shown is always for a single session; however the masterpoint award may be based on performance over multiple sessions. For multiple session events, players receive the greater of the (1) sum of the masterpoints received for their session ranks; (2) masterpoint award for their average percentage. ACBLscore computes this but does not have enough room on its old style reports to fully display the logic behind multi-session awards. Thus a large masterpoint award in the afternoon session for a mediocre percentage usually means the pair had a very good morning session and their average percentage from the two sessions ranked well.

4. Face View and Masterpoint Tooltip

If your ACBLmerge administrator has set up the Face View feature on the web server and selected it for a given report, holding the mouse cursor over a player’s name in the recap section of the report will bring up a picture of the player if a picture is available. If masterpoint display has also been selected, the player’s approximate masterpoint total will be shown.

Uncheck the “Show Faces” box to turn off the face display. In the iPhone customized report, face display is turned off initially because of the small size of the display. Check the box to enable it.

6. The Board

The board shows the board number, the board hands, the dealer, and vulnerability just like a standard recap sheet. A dash indicates a void suit. ACBLmerge uses standard font characters for the suit symbols, so you can readily copy-and-paste a single hand. If you try to copy-and-paste the entire board, you may not get what you expect; this is explained further in the Copy-and-Paste section below.

7. Results for each Pair

The raw score (e.g. 140), and the matchpoints (e.g. 16.00) are shown and sorted in descending number of matchpoints for N-S, an idea borrowed from Evan Bailey. This presentation makes it easier to see how often each result was achieved and where you stand on the board. N-S players want to be near the top and E-W players want to be near the bottom.

If Bridgemate, BridgePad, or Bridge Scorer electronic scoring results have been included, the contract and optionally the opening lead will also be shown as illustrated by item 6(b). Most electronic scoring systems do not check if the opening lead entered into the system belongs to the hand that was on lead and therefore allow data entry errors. ACBLmerge highlights erroneous leads. Yellow highlighting means the lead card was in hand of opener’s partner; red highlighting means the lead card was in declarer or dummy’s hand. Leads highlighted in yellow are often correct, the actual problem being that the contract was recorded as being played from the wrong side. Sometimes this happens because the board is play upside-down and nobody notices. Less commonly, it indicates a lead out of turn that declarer accepted. Leads highlighted in red are clearly erroneous. Often, a player just entered the wrong value of a low card, e.g. ♠2 instead of ♠3.

To find where you are on each board you may use the search functionality in your browser. Ctrl+F will bring up a search dialog in all main web browsers (use ⌘+F instead on Apple computers). Just type your last name and press Enter. To move to the next board in IE press, Enter again or the Next button to repeat the search. In Firefox or Chrome press Ctrl+G (⌘+G on Apple computers) to repeat the search or simply press the Highlight All button on the Find bar. Alternatively you can open a recap sheet for your partnership.

For those unfamiliar with matchpoint scoring, you receive one matchpoint for each pair you beat and half a matchpoint for each pair that you tie. For a board played N times, (N-1) is the best possible matchpoint score, a “top”. Unlike team play (IMPs), the relative difference in raw scores does not matter. Thus if 3NT and 4 not vulnerable both make four, then the pairs scoring +430 (3NT+1) beat the pairs with +420 (4) to the same degree that a pair making slam beats a pair that fails in slam. Then again if 3NT makes three and 4 makes four, the +400 loses to the +420.

If a board is played fewer times than other boards, for example due to the movement or because it was pulled from a slow table, ACBLscore will scale up the matchpoint values, leading to funny values which are not integers or half integers. Scaling also occurs in multi-section events with an unequal number of tables. Director rulings that award AVE+ (60%) and AVE- (40%) to the non-offending and offending sides respectively can also lead to strange looking matchpoint scores. Rare split rulings where the non-offending side receives an AVE+ and the offending side a separate score are an additional complication. Split ruling are indicated by a vertical bar, e.g. AVE+|-800 means your side received an AVE+ and your opponents received -800.

8. HCP total for each hand

Shows the high card point (HCP) for each hand, based on the standard A=4, K=3, Q=2, J=1 point count system. No adjustments are made for length and/or shortness.

9. Double Dummy Makeable Contracts

Shows all contracts that can be made if all players knew all four hands and bid and played perfectly. Since real bridge players do not see all the hands and therefore must bid and play based on the ensemble of hands consistent with the bidding and play up to that point, double dummy results must be taken with a healthy dose of skepticism. The double dummy line of play may require dropping an offside singleton king, underleading an ace, eschewing a simple 50% finesse for a low probability squeeze requiring exactly the right lay of the cards, and other plays that work on the given board but which would not be the percentage bid or play. Many new players look at the double dummy results and feel they have done something wrong if they do not match the computer generated result. This is a serious misconception. To some extent, double dummy results are shown because double dummy analysis is a well defined and solvable problem, requiring less than a second on average for each board running on a modest computer using efficient software. Determining what should happen on a board is much more complicated and even then would involve assumptions about bidding systems, aggressiveness, and so forth, which vary.

The ACBLmerge double dummy results are complete. Every denomination is tested from each seat. By contrast the Deep Finesse results shown on Dealmaster generated recap sheets are usually incomplete. In the default setting, Dealmaster only computes the double dummy results for “likely” contracts. Sometimes it misses a good, or even optimal, contract. Usually this happens when the trump fit is poor, e.g. a 4-3 or 5-1 fit with strong cards. Thus, a dash on a Dealmaster recap sheet does not always mean the contract fails.

On the ACBLmerge output, a result such as 4♠ means that the contract can be made regardless of direction. A result like 3/4 means North (or East) can make 3 and South (or West) can make 4. A dash, as in -/1 means only South (or West) can make anything in hearts. Smaller indications in light blue such ♠5, with the number after the denomination, show the number of tricks that can be taken. Thus, ♠5 means 2♠ would fail by three tricks. Sometimes one actually cares to know just how many tricks a low level contract can be set. For example suppose you are vulnerable against not and the auction proceeds 1♣ 1♠ to you. Holding ♠KJ83 AT5 KQ72 ♣J4, you debate passing and hoping that partner reopens with a double, but decide to bid 3N directly instead. After you collect +660, you begin to wonder if defending 1♠* would have been more profitable. ♠3 would tell you that +800 had been possible.

If after a bit of thought, you are not sure what lines of play the computer has run through to reach its double dummy conclusion, you can load the board into a double dummy analyzer which will allow you to see the double dummy result for any starting line of play. My favorite Windows program for this is Bridgify. It is free, has a clean, simple, intuitive interface, analyzes hands quickly, and can load hands from a PBN file. Conveniently, it is a standalone executable and therefore does not need to be installed. If you copy the executable bridgify.exe and help file bridgify.html on to a USB memory stick you can take it with you and run it from any Windows computer. The free Bridge Captain Double Dummy Solver is another alternative for Windows. Its interface is not as clean but it has the ability to compute the par contract and show how many IMPs were won or lost on each hand relative to the par contract.

If your club’s webmaster has uploaded the PBN file, you will be able to download it via the link “Download hands in PBN format” at the top of the ACBLmerge report. Then in Bridgify you can load a board via Load (button) → Open File (button) → {choose PBN file} → Open (button) → Choose hand (list box) → Load (button). If the PBN link is not provided or is broken, click on the C button at the bottom of the hand to bring up the Cut-and-Paste Aid. The “Bridgify ready format” lists each hand in the format one enters them into Bridgify after clicking the New button in Bridgify.

10. Law of Total Tricks Calculation

The so-called Law of Total Tricks (LoTT), formulated by French bridge theoretician Jean-René Vernes in an article in The Bridge World in June, 1969 and popularized by Larry Cohen in his books, To Bid or Not to Bid (1992) and Following the Law (1994) is a theory that states:

“The number of tricks your side can take, if you play in your best trump suit, added to the number of tricks the opponents can take, if they play in their best trump suit, is approximately equal to the number of trumps held by both sides.”

The LoTT is known to systematically underestimate the total number of tricks when the number of total trump is high, there are double fits, or very pure suits. The final chapter of To Bid or Not to Bid discusses adjustments to make. Even with adjustments, the LoTT has also been criticized as simply inaccurate, most notably by Mike Lawrence and Anders Wirgren who wrote I Fought the Law of Total Tricks (2004). Nevertheless, the LoTT is often a useful guide for competitive bidding decisions.

The ACBLmerge report shows the number of total tricks, the number of total trump, and the difference between the two. If a side has more than one longest suit, the suit yielding the highest number of tricks is considered, the standard LoTT interpretation. If the LoTT were 100% accurate, the final number would always be zero. LoTT fans can see if they can justify the adjustment implied when the difference is not zero. Detractors can gloat when even with adjustments, the LoTT is significantly off.

11. Par Score and Contract(s)

The par score is the raw score for the contract beyond which neither side can bid further to improve the score for their side. The par score is shown from the perspective of N-S, i.e. it is negative if E-W are making a contract or N-S must sacrifice. There may be multiple par contracts which achieve the same score. For example, equivalently scoring contracts such as 4 and 4♠ will yield the same par score. Likewise 3N and 5♣/; however if the opponents can sacrifice profitably in say 4♠ but not 5♠ then only 5♣/ would be the par contract. All sacrifice contracts are doubled since that always improves the result for one side. As with makeable contracts, there may be equivalently scoring sacrifices in more than one denomination.

The par score calculation is computed from the double dummy information above and therefore carries the same caveats. Accordingly, some authors speak of the par presented here as “absolute par”, reserving the word par for the most realistic result given what each player knows as the hand is bid and played out.

Although par is a straightforward concept, there are some subtleties which the code must handle. First and rather rarely, neither side may have a makeable contract in which case the par score is 0 and the par contract is Pass Out. Thomas’s Bridge Fantasia has an interesting section on these par zero deals. Second, and somewhat more commonly, both sides may be able to make the same maximally scoring contract (often 1NT) from at least one seat. In game theory, this is a “hot” situation where victory goes to the side to get there first. This situation is indicated by the ± sign, e.g. ±90 1NT-EWNS. See Everyone makes 1♠ and Everyone makes [3N] for specific examples. Victor Mollo anticipated such amusements in Bridge in the Menagerie, where Oscar the Owl, his chief kibitzer, remarks, “Curious hand, both sides can make four hearts”, one by drawing trump and establishing a side suit, the other by cashing outside winners and then crossruffing. However, 4 is probably not the par contract on that hand.

12. Copy-and-Paste Aid

Copying and pasting from a web browser is a treacherous affair. The results depend on your operating system, the web browser, and the target application, e.g. an e-mail program or Microsoft Word. If you copy a single hand from a board or copy the pair results, you will probably get what you expect. If you copy the entire board, you may not get what you want, especially in Firefox which seems to strip away all the formatting supplied by Cascading Style Sheets (CSS); IE7 and IE8 seem to get this mostly right. Moreover, you may wish to have one or more hands in horizontal format or a pair of hands side by side in vertical format. To accommodate these needs, the copy-and-paste aid button will launch a pop-up window which presents the hands in various formats, ready to be copied and pasted as needed. It is impossible to anticipate every need, so please e-mail me if there is a format you desire.

Sophisticated users might ask why ACBLmerge does not provide a way to directly copy the exact HTML onto the clipboard at the press of a button. This because allowing web applications to write directly to the clipboard is a security risk. All serious web browsers either do not allow it at all or make it very difficult and even then still prompt the users to make sure they really want the data placed on the clipboard. Therefore, to accommodate a variety of operating systems and web browsers, it is best to simply have the user actively copy data to the clipboard.

It is possible that the Copy-and-Paste Aid popup window will be blocked by a pop-up blocker even though most pop-up blockers are now smart enough to distinguish between user launched pop-ups and automatically launched pop-ups. If this happens, you will need to edit your pop-up blocker settings or tell it to allow popups from lajollabridge.com or whatever website is hosting the ACBLmerge results. Also, if you save a results page to your local filesystem on Windows and open it from there and try to launch the Copy-and-Paste Aid from Internet Explorer 7 or later, you will probably receive the message “To help protect your security, Internet Explorer has restricted this webpage from running scripts or ActiveX controls that could access your computer. Click here for options...” If you want to enable the Copy-and-Paste window to work from a local file, you will have to change the setting accordingly.

13. Team Games

For team games, holding the mouse cursor over a team number or an opposing team number, i.e v##, will show the members of the team, the average and geometric average of the masterpoints held by each team member, and the team’s seed rank. The seed rank is generated by ranking ordering the geometric mean of the teams masterpoint holdings. This seed rank does not necessary match how the ACBL or WBF performs seeding in events that are seeded. It merely reflects my approach though it seems reasonably accurate when compared to outcomes.