Join Bridge Winners
Bridgescore+ used for Truscott Senior Swiss Teams
(Page of 5)

Bridgescore+ ran the first day of the Truscott Senior Swiss on Monday.

These are my notes/comments/observations.

The event started on Monday at 1pm. 7 board matches, 4 matches (rounds) in first session, 4 in second session. Second session at 7.30. Qualifying for Tuesday. The first day (both sessions) was run using Bridgescore+. As I was playing on Tuesday in a different event, we decided before the tournament that Bridgescore+ would be for one day only. A good opportunity to show case new technology - thank you Joe, Gary, Greg, Tom, Steve from ACBL.

I had some equipment stored in an ACBL room so picked it up around 11 am. Took 10-20 minutes to get and set up.

All assignments were up before 12:00. Play started at 1pm.

The goal was to run the event, as well as some real world training for the TDs.

A TD set up the event around 11:15, configured the table inventory, expected number of entries (100-120). He set up the first round, but had to leave by 12:00 for another event.

We had a stack of entries from Bridge Winners. These were randomly assigned team numbers in the 1-80 range. I picked them up around 11:00. Cut off was 10 am.

There were 3 projectors in use; one powered with a Raspberry Pi, the other two with ACBL laptops. There were 3 input stations (places where players entered scores). One projector was mine; two were ACBL. Mine was an older model; ACBL had short throw projectors which are better as they can be closer to the wall. The only important thing about a projector is that players do not trip over a cord. You would not believe how inventive players are about trying to walk over the cord. Anything that blocks a cord - chairs, sharks - is considered a challenge that must be overcome. We even tried police tape at one tournament but that was a magnet. Players - please, please don't walk over an area blocked off. Also, if you have to stand so close to the wall that you block the light from the projector, please don't complain when you can suddenly no longer see it.

The input stations were laptops. Those with a 10 key are better. One had an external 10 key.

Teams were given 58 minutes to play each round. I decided to use the local version of Bridgescore+ not the cloud version as had heard of limits on the number of connections on the WiFi provided by the hotel to ACBL. Apparently the limit was reached earlier in the tournament. I didn't want any such problems with running a National event.

The DIC, who is nameless but for his initials KVC, made the initial announcement that this was a timed event, 58 minutes long, and that the first time was a warning; the second time a penalty. The time limit is for scores to be reported. 7 boards matches, 8 minutes per board (including shuffling time), 2 minutes to work out scores and report. 58 minutes. BS+ can track exactly when an assignment is posted, and exactly when a score is posted. Unknown to the players is that the decision was made, certainly for the first session, that there would be no late penalties (unless egregious) because of the new technology. I think this was the correct way to handle this.

BS+ can track the late plays; but assigning penalties is the responsibility of the TD. There were no late plays penalties assessed. There was a ZT penalty assessed. 

Round 1 went well. But as soon as players started entering scores, the posted VP was wrong. ACBL were using the WBF 20 VP scale with decimal points. The problem was the event had been configured to run 5 board rounds, but 7 were in play. A simple error in configuring the event. There is a different scale for 5 board rounds and 7 board rounds. After round 1 was over, the event was reconfigured, the scores were corrected using the correct formula for 7 board matches. A slightly embarrassing start, but it did not affect matching for round 2. All subsequent scores were correct. A player first reported the problem. We had done lots of testing, so I knew it was not a bug - I suspected the issue but decided to let all scores in round 1 be posted.

A bigger problem in round 1 was stationaries. There were a large number of stationaries (16) for a 108 table event. But... this IS the Senior Swiss. Plenty of testing had been done with large number of teams and large number of stationaries. There were some needed stationaries, e.g. wheelchairs, power requirements. There were some desired stationaries - players that prefer to remain stationary.

Some people had bought their entries on Bridge Winners and requested stationaries. Some people bought their entries at the door and requested stationaries. There were multiple sellers and one team was sold a stationary at C23. Another entry was sold a stationary at D23. We had a manual list of stationaries and unfortunately no one noticed the problem for round 2. Hopefully you can.

The assignments were going up for round 2 but some teams were not being assigned. I saw the issue, but was not exactly sure of the problem at the time. I made the decision to remove the stationary flags from all teams, except for those that I knew were really stationary teams, e.g. required power or wheelchair. Assignments were then made for round 2. I knew that this meant that some (most) of the stationaries would have to move; but as I was not sure of the real problem this seemed the best interim solution. Trying to find the problem did not seem the correct solution as I was not sure how long it would take. If you know of a timely workaround, you do it, then deal with the fall out later - this is what happened. I didn't work out what the original problem was until some time into round 2.

Unfortunately... with this number of teams and the large number of stationaries, it was inevitable that in round 2 some stationary teams - that had just been moved - would be paired against other stationary teams that had just been moved. And this is what happened. So... more teams would have to be moved in round 3. Yuk. 

By Round 3, we knew exactly what teams were stationaries. We had moved one team with a player in the wheelchair that was originally assigned to the far end of the room, behind a pillar to a table by the door (sorry - you should never have been assigned that table in the first place). That was a deliberate move for Round 2. Others were not.

In addition to the original C23/D23 there was another set of tables that had been similarly configured (I forget the exact table number).  These were also fixed.

By the end of round 3, all of the stationary tables issued had been fixed. But we had the issue where teams that had requested stationaries, had now been asked to move for 3 rounds. That's unacceptable and quite rightly, some people were complaining

I can only hope that by explaining what went wrong, that there is some understanding of the issues.

For the second session, we were able to configure all stationary table requests so that they were all close to the entrance. For those that needed power, they were at an assigned table. For those that requested superior lighting, they were accommodated.

The second session went smoothly; no stationary teams had to move. This is what it should have been for the first session, so for all affected teams, please accept my apologies.

We configured the table inventory so we used the tables that we wanted the players to use. We kept them away from the main TD table.

So, what went wrong? 

We sold at least two teams (I think there was 4) stationary tables, told them they would be stationary for the event, and they played each other in the first round. We failed to prevent this, and also to detect this. I use the term 'we' because it doesn't matter if it is a people/process or software issue - the players don't care - they only know there was a problem.

So, how do we prevent it from happening again?

What should have happened is that the sellers should have been given a set of stationary tables to assign, and held those tables for the stationary requests. When someone requested a stationary table, they should have been assigned to one of those tables. This is a process related issue, and something that is easy to fix and implement. For future events with BS+, I will make sure this happens. I have already added this it to the printed checklist for BS+ Swiss events.

In my view, running a Bridge event is like piloting an airplane. If you follow check lists, you will not forget anything.

BS+ suggests the use of a sign-up sheet to detect these types of problems. We use it in Gatlinburg for the large KOs. For some of these KOs, we actually have to use two sheets because the number of requests fills more than one page.

We failed to use this form for the Truscott Swiss. This was my fault in not providing it to the sellers and in training them in how to use it. This is also a process issue.

The software needs improving to detect this type of configuration error, and flag it in round 1 so it can be corrected early. This is something I can fix. During testing - surprise, surprise - we never thought this error could occur. I'm sure it won't be the first such issue.

Other issues:

We had one team that misread the assignments on the projector. Their last digit ended in 0 and they read the team number that ended in 8 instead. An honest, reasonable mistake. They started to play, only for the real team to come to the table. This is an easy problem to fix. We have a tablet that TDs can walk around the floor with the list of assignments. This problem also occurs with the rack, it is easier to fix if the TD has the assignments in front of him.

Remember - this is the first time that many of the players had ever seen the screen. The first time that the players had ever been asked to enter their own scores. I was expecting some teething problems with player education. The display screen is now in its 5th iteration - each iteration better than previous but it is still a new screen for players.

I keep using Gatlinburg as an example, but there are reasons - there are few events this large. In Gatlinburg the floor TDs will take a picture of the assignments screen on their phone, so if there are any teams in the wrong place, the TD has the information right in front of them. This is an example of fixing a problem multiple ways. But... until TDs (and players) have more experience with Bridgescore+, these learning curves are expected.

Several ACBL TDs have used Bridgescore+. I try and be hands off when running an event. TDs learn by doing, not by watching someone do it. For the Truscott, there was only 1 TD that had seen the software before, and he was there to help set up the hardware. So... National Event ... no TDs that have used the software before. Please note that I was not in charging of scheduling. I have always claimed that BS+ is simple enough to use that you don't need much training, and ACBL decided to put this to the test in a National Event. 

Training was during the 1 hour before the event. The TDs all had to be trained in how to explain to players to use the "Player input stations". All of them had to be trained in what to do when a team enters a wrong score. This will happen a couple of times per session - usually entering the score backwards. There is a simple fix; but the fix is more complicated if matches have been assigned.

For the player name entry, within a couple of minutes of the TDs arriving, and given basic instructions, they were entering player names, stationary tables etc.

The "player input stations" are new to the players. For those not there, there is a web based interface where players enter their team number, hit enter. It displays their opponent number and asks for the IMP score. You enter it, with a minus if you lost, then hit enter. The score is sent to Bridgescore+.

We set up 3 player input stations. One on each of the three sides of the room (not by the door). The interface is simple enough that the players need no training. There was a TD assigned to each station for the first 3 rounds. The players seemed to have no problems entering the scores. We put them by the projector because this is the logical place to return scores (where they are posted!).

BS+ uses the published ACBL Swiss Matching algorithm. Unfortunately ACBLscore uses something slightly different. If there are 3 teams that have the same score, the decision on which two to match is random; therefore BS+ and ACBLscore will never be the exact same. However, BS+ posts assignments earlier than ACBLscore so I know the ACBLscore algorithm is different to ACBL. This may have affected some of the assignments but not so that anyone would notice. I'm going to try and get the actual ACBL algoirthm, not the published one. BS+ implements both the ACBLscore method and also perfect matching.

Automatic matching was turned off during Round 4. After all scores were in, perfect matching was used for Round 5.

After the first session, two teams failed to report. I was told to enter a tie; then get the result during the next round.

BS+ has a feature where you can configure a team to "See TD" rather than display their assignment. Those 2 teams had to come find a TD before they could find out where they were playing in R5.

After the correct result was entered, a different TD told me that the correct procedure is to award one team a win of 10. Assign using perfect matching. Then find out what team won. If you guessed wrong, swap the team assignments for round 5. Using a "win by 10" value works out better than "assign a tie" as easier to fix teams. I'll add that to the BS+ procedures.

At round change, the TDs have little to do. Assignment is automatic. Tabling is automatic. During round change, I quietly sat behind the TD table.

The first session had hiccups, and some learning curves for TDs, players (and me).

The second session was, IMHO, very smooth.

I have always claimed that BS+ is relatively easy to learn. ACBL put it to the test with a National Event and no training for the TDs. Outside of the stationaries issue, I think it worked very well. 

For the second session, the players knew how to enter the scores.

Over half the field were done within 4 hours for the second session. This is about 15 minutes earlier than similar events earlier this week with fewer teams.

I packed up my stuff and walked it back to the storage room. At 12:15am, 9 cleaners from the hotel came in to the room. They were there for about 30 minutes. The things that players never see.

I believe it is important to share what worked, what didn't, and why, and that any problems are being addressed. 

I did get some feature requests. For example, display the current qualifying value so teams know what to aim for (it's nearly always half the maximum score + 1).

Also, highlighting teams that qualified at the end of the event.

These are all small items which will help improve the product and will be in the next release.

If anyone else has any comments/suggestions, please let me know. PM is fine.

Again, I apologize for the stationary team problems. Hopefully I've explained what happened, what we did to fix it, and why it should never happen again.

28 Comments
Getting Comments... loading...
.

Bottom Home Top