[criptik]

Occasionally we have had situations where there is more than one robot pair in our small unclocked tournaments. (One to fill a sitout, and one to fill humans who had to leave early). It seems to work fine, but a few questions:

• When you look at a traveller that has a robot pair in a tournament with 2 robot pairs, is there any way to distinguish which robot pair you are looking at?
  
  
• In our howell-like movement, eventually the robot pairs have to play each other. I've noticed the completion times for each board in such a robots vs. robots round is the same, so the hands essentially take zero time to play (and of course the robots then wait a good while for some human player to finish). I assume this is normal?
  
  
• When two robot pairs sit the same direction in a round (say as declarer), it looks like they don't always play the same cards in the same situation. Is this normal?

[m00036]

criptik, on 2020-August-22, 17:06, said:

Occasionally we have had situations where there is more than one robot pair in our small unclocked tournaments. (One to fill a sitout, and one to fill humans who had to leave early). It seems to work fine, but a few questions:

• When you look at a traveller that has a robot pair in a tournament with 2 robot pairs, is there any way to distinguish which robot pair you are looking at?
  
  
• In our howell-like movement, eventually the robot pairs have to play each other. I've noticed the completion times for each board in such a robots vs. robots round is the same, so the hands essentially take zero time to play (and of course the robots then wait a good while for some human player to finish). I assume this is normal?
  
  
• When two robot pairs sit the same direction in a round (say as declarer), it looks like they don't always play the same cards in the same situation. Is this normal?

• Frustratingly no (as far as I can tell). When 2 robot pairs are required, I tend to record the movements of each robot pair and their opponents so that they can be matched up (Mitchells and Howells at least have clear movements but table numbers are not recorded after the event so you have to do this in real time). You can also trial and error so that the percentages line up of course. It may be that there is a way which I haven't found yet though.
  
  
• Yes robots will finish the round in next to no time.
  
  
• They will always play the same cards in the same situation but the crucial part is that the bidding may be subtly different, meaning that the situation is indeed different. Robots assume their opponents are playing their system, so even something as simple as a 1NT open showing 15-17 at one table and a return to 1NT (showing 1NT but robots interpreting as 12-14) could affect the defence as the probability of declarer having certain cards will change. It helps to explain why robot defence can be so good at times and so frustrating at other times (e.g. the robot will hardly ever play you for points you "can't" have if there is a "legitimate" line of defence available, but it may just be that the robot and the humans are playing different systems).

[barmar]

criptik, on 2020-August-22, 17:06, said:

Occasionally we have had situations where there is more than one robot pair in our small unclocked tournaments. (One to fill a sitout, and one to fill humans who had to leave early). It seems to work fine, but a few questions:

• When you look at a traveller that has a robot pair in a tournament with 2 robot pairs, is there any way to distinguish which robot pair you are looking at?
  
  
• In our howell-like movement, eventually the robot pairs have to play each other. I've noticed the completion times for each board in such a robots vs. robots round is the same, so the hands essentially take zero time to play (and of course the robots then wait a good while for some human player to finish). I assume this is normal?
  
  
• When two robot pairs sit the same direction in a round (say as declarer), it looks like they don't always play the same cards in the same situation. Is this normal?

1. No. We have internal names that distinguish them in the database, but all these names just get translated to "Robot" when showing results to humans.

2. Yes. If you've ever played a robot game without the "human declares" option, and the robot opposite the human becomes declarer, you'll see that the hand is played at lightning speed. The only thing slowing it down is the browser's display of the cards being played, but that doesn't exist when the robots are playing each other in a tournament.

3. I haven't checked, but I wouldn't be surprised. In robot duplicate tournaments we ensure that all the robots are using the same random sequences so they play the same at each table. But we may not do the same thing for the robot substitutes. So there's a little extra randomness there.

[criptik]

barmar, on 2020-August-26, 18:03, said:

1. No. We have internal names that distinguish them in the database, but all these names just get translated to "Robot" when showing results to humans.

It's a very low priority request but it would sometimes be nice for the humans to distinguish the robot pairs in the travellers.

[pescetom]

criptik, on 2020-August-27, 12:37, said:

It's a very low priority request but it would sometimes be nice for the humans to distinguish the robot pairs in the travellers.

Please make that a medium priority request to name robots with their id rather than just 'robot'.
One case in which this is genuinely needed is when a player drops out because of network problems and is substituted by a robot, but then reappears and wants his seat back - if there were already other robots playing in the tournament then director has no obvious way to know which to substitute.
I am thinking in particular about individual tournaments, but even in a pairs tournament he might not have a trace of original pairs.
Of course this would not be a problem if the whole problem of dropout replacement/reentry was automated, which is a high priority request

[m00036]

pescetom, on 2020-August-27, 15:01, said:

Please make that a medium priority request to name robots with their id rather than just 'robot'.
One case in which this is genuinely needed is when a player drops out because of network problems and is substituted by a robot, but then reappears and wants his seat back - if there were already other robots playing in the tournament then director has no obvious way to know which to substitute.
I am thinking in particular about individual tournaments, but even in a pairs tournament he might not have a trace of original pairs.
Of course this would not be a problem if the whole problem of dropout replacement/reentry was automated, which is a high priority request

The problem of automatic replacement/reentry is that not all clubs will want it so it has to be added in as an option for TDs which takes extra time to implement. My clubs would not want such a feature but I know in larger BBO tournaments it works very well so depends on the situation.

Anyway on the point of the thread, you can use the Reseat Player command to sub players back in without having to know where they were before.

[diana_eva]

Reseat player is how to put the player back. This is available in all tourneys, not specific to pay ones.

[pescetom]

diana_eva, on 2020-August-27, 16:54, said:

Reseat player is how to put the player back. This is available in all tourneys, not specific to pay ones.

Ok thanks, I missed it and was using substitute with select.

[pescetom]

m00036, on 2020-August-27, 15:52, said:

The problem of automatic replacement/reentry is that not all clubs will want it so it has to be added in as an option for TDs which takes extra time to implement. My clubs would not want such a feature but I know in larger BBO tournaments it works very well so depends on the situation.

Obviously it should be a choice, just like allowing robots and so on. Making it a choice costs almost nothing in implementation terms and is no stress for directors not interested (just another tickbox on the 'entries' tab of 'create tournament').