Join Bridge Winners
ACBLscore Bug

Tournament directors (TDs) around the country are calculating masterpoint awards by hand this week because of bugs in the most recent update to ACBLscore. The bug involves changes to masterpoint awards that went into effect for 2015, and was noticed only a few days before the end of 2014 by a diligent TD who was preparing for the regional in Charleston, SC, which ran from Dec. 29-Jan. 4. In addition to tournaments, the bug is apparently affecting clubs running Junior Fund (and perhaps other special) games.

The formula for masterpoint awards is always being updated. This is the responsibility of the ACBL Board of Directors (BOD). A number of changes for 2015 were approved at the BOD summer meeting in Las Vegas, and then given a second reading and final approval at the fall meeting in Providence. Among the changes approved are

  • Increasing the match awards in the Spingold, Vanderbilt, Wagar, and Senior knockouts.
  • Tweaking the masterpoint awards for the strats of stratiflighted pair games. (This was necessary because the BOD approved changes that allow six strats in a single game, meaning a strati-flighted game can have three flights with two strats each.)
  • Increasing the “depth” of overall awards. (Meaning more pairs/teams can get overall awards.)
  • Tweaking knockout payments.
  • Increasing payouts in some NABC events, including the NAP and GNT.

The BOD met in Providence before the NABC and approved the changes to the masterpoint formula; the update to ACBLscore incorporating these changes has been in place since November. Everything went fine until the calendar hit January 1 and the new masterpoint calculations took effect. According to the TDs I’ve spoken to, some of the expected changes aren’t there, and some have caused huge errors. Events with a stratover 3,000 masterpoints are reportedly generating an overall award of 0 masterpoints, and very few (if any) masterpoint awards being generated are correct.

The Director-in-Charge (DIC) of the Charleston regional was apparently the first to notice the problem. Knowing that his tournament included events in both 2014 and 2015 and that the masterpoint rules were changing, he decided to do some testing before the tournament began. This was around December 26. Using some old and made-up data to simulate events in January 2015, he ran ACBLscore and noticed problems with the masterpoint calculations.

He informed the ACBL, but they were unable to get the bugs fixed by January 1. So he and other ACBL employees came up with work-arounds to get the masterpoints correctly calculated manually by the TDs at the tournaments until the software can be updated. The ACBL sent the DICs of the early-January regionals an Excel spreadsheet to help them calculate masterpoint awards (particularly the new overall awards) correctly. So far regionals affected have been the ones in Charleston; Newton, MA; Orlando, FL; and Monterey, CA.

There are several questions raised by this debacle. Why wasn’t the new release of ACBLscore tested? Some of the errors it is generating are so obvious (paying 0 masterpoints!) that any rudimentary testing would have caught them. The ACBL has a poor track record in this area. TourneyTrax was apparently a disaster when first released. The new website had issues. ACBL Live was sorting dates improperly, so the first results being shown were from 2013. Why doesn’t ACBL test software before releasing it?

Even without testing it themselves, they were given nearly a week’s notice; why wasn’t the ACBL able to fix the problems and get a new version of ACBLscore out by December 31? The answer to that one is simple: Jim Lopushinsky (affectionately known as “Lopo”) was on vacation. Lopo is the original programmer of ACBLscore and has been responsible for its maintenance since the 1980s.

One of the major concerns about ACBLscore, and a driving factor in the decision to build a modern replacement, was the dependence it created on one man. Many people have voiced this basic concern to me: If Lopo were to die (or quit), what would happen to ACBLscore? When I talked with the newer members of the ACBL IT staff, who recommended abandoning the ACBLscore+ project and focusing instead on updating the current ACBLscore, I naturally brought up this concern: why would you recommend updating this program that no one but Lopo can access? They told me that one of the side benefits of the ACBLscore+ project was they were able to “open up” ACBLscore and realize how much they could actually do with the program. Without Lopo.

But now something breaks, and they can’t fix it without Lopo. Does that mean they can do new stuff (like ACBL Live) without Lopo, but he is still necessary for basic maintenance? That still makes the ACBL dependent on one man, who has to be near retirement. What is the plan for a few years down the road if Lopo retires? How can the ACBL justify the risks associated with having the fate of its key piece of software — the link between headquarters and all the clubs and tournaments, the thing responsible for determining and issuing the ACBL’s currency, masterpoints — tied to one programmer?

For now, the TDs are doing yeoman’s work getting masterpoints calculated, and if you are at a tournament this week please tell them how much we appreciate them and the work they do. Presumably when Lopo gets back from vacation the bugs will get fixed, a new release of ACBLscore will come out, and things will go back to normal. But serious issues have been raised, both about the ACBL’s methodology (i.e., lack of testing) and its forthcomingness about the actual dependence on Lopo going forward.

92 Comments
Getting Comments... loading...
.

Bottom Home Top