[jtfanclub]

Hi all, been trying my hand at running tourneys, and for the most part they're doing well. Some software stuff I'd really like to have are:

1. When I hit the button to kick people out of the queue before the tourney starts who aren't online, it would be really appreciated if it would also kick out people who are currently playing in other tourneys (usually because they subbed in). I'm tired of having 36 pairs signed up and getting somewhere between 14 and 17 tables. People who are offline might come back. People who are playing in another tourney aren't.

2. I really really want to be able to set the minutes per board with a decimal points. For example, my tourneys have 6 minutes per board, 4 boards per round. When time is running out, usually about a third of the people are still playing the last board, which is no problem...I'm happy to give them a couple of extra minutes to finish. But there's also almost always one or two tables still playing the next to last board, and I don't want them starting the last board and delaying the entire tourney for 6-8 minutes. Unfortunately, if on minute #23 I change the tourney from 6 minutes per board to 7, and then on minute #24 somebody finishes the next to last board, it lets them start a new board. If I could change it to 6.5 minutes per board, it wouldn't allow people to start new boards while giving the people on the last board an extra two minutes to finish.

3. I'm probably the only person on the planet with directions issues. Half the time, it seems like "They" is EW while half the time it's NS. I'm also very unused to looking at history and seeing it in SWNE format, since it seems like most other places the partners are together. This becomes a problem when I'm adjusting a board, especially due to a sub (so it's not obvious who was playing with who).

4. Can I have Victory scoring, please? 1 point for winning a round (using IMPs), 0 points for losing, .5 for a tie. Highest total wins. The problem with both IMPs and MPs is that they can become drunk-punching contests: if you get a 100% the first round and a 40% the second, it's a better result than the pair that got 60% in both. I'd like to run KO tourneys, in which the winner goes on and the loser switches brackets or gets kicked out, but as it's currently set up the loser might go on and the winner get kicked out if the loser had a higher margin of victory the previous rounds. 20 or 30 point VP scoring would also work, but give me the 0-.5-1 any day.

5. When I asked for somebody to be subbed, I'd like it if they got well and truly booted. I had both sides of a partnership being problems at one point, and each was refusing subs for the other.

6. Besides accept and decline director calls, could I get a "hold on a moment" option that will both send that message to the person and make the director call go away only to come back in one minute? It's really annoying when I'm trying to fix somebody only to have to tell the people that also need help that I'm not ignoring them on purpose.

OK, I think that's it. Sorry if some of this already exists and I haven't found it. A million thanks to all of you for this place, and it's a lot of fun to run tourneys here!

[mink]

hi jtfanclub,

Quote

2. I really really want to be able to set the minutes per board with a decimal points. For example, my tourneys have 6 minutes per board, 4 boards per round. When time is running out, usually about a third of the people are still playing the last board, which is no problem...I'm happy to give them a couple of extra minutes to finish. But there's also almost always one or two tables still playing the next to last board, and I don't want them starting the last board and delaying the entire tourney for 6-8 minutes. Unfortunately, if on minute #23 I change the tourney from 6 minutes per board to 7, and then on minute #24 somebody finishes the next to last board, it lets them start a new board. If I could change it to 6.5 minutes per board, it wouldn't allow people to start new boards while giving the people on the last board an extra two minutes to finish.

This is a problem only if you have rounds of 3 or 4 boards. I do not like such tourneys, and I tell you why: In order to have a tourney that it as fair as possible the average skills of the opps of any given pair should be as close as possible to the average skills of all pairs of the tourney. The odds to achieve this increase if you increase the number of opps for a pair. This is done by decreasing the number of boards per round. So 1 board per round tends to be as fair as possible to this respect. However, with 1 board per round you cannot compensate time lost at the first board by playing the second faster. If you think this is an issue, 2 boards per round is a compromise. Another argument for few board per round is that the annoying breaks between the round for fast tables are longer if you have more boards per round.

So I think it is not worth investing work here for a function that is needed only by few for a questionable purpose.

Quote

3. I'm probably the only person on the planet with directions issues. Half the time, it seems like "They" is EW while half the time it's NS. I'm also very unused to looking at history and seeing it in SWNE format, since it seems like most other places the partners are together. This becomes a problem when I'm adjusting a board, especially due to a sub (so it's not obvious who was playing with who).

This is easy: When you play we and they is correct. And when you kibitz or direct, we refers to NS and they to EW. Always.

Quote

4. Can I have Victory scoring, please? 1 point for winning a round (using IMPs), 0 points for losing, .5 for a tie. Highest total wins. The problem with both IMPs and MPs is that they can become drunk-punching contests: if you get a 100% the first round and a 40% the second, it's a better result than the pair that got 60% in both. I'd like to run KO tourneys, in which the winner goes on and the loser switches brackets or gets kicked out, but as it's currently set up the loser might go on and the winner get kicked out if the loser had a higher margin of victory the previous rounds. 20 or 30 point VP scoring would also work, but give me the 0-.5-1 any day.

This is essentially what you get if you have a 2-table-tourney with matchpoints scoring, or a "board-a-match" team fight. But if you have more tables, how do you define winning and losing? E.g. with 4 tables and 4 different results, what do you score for the middle results? I cannot see how this can work for more tables.

Quote

5. When I asked for somebody to be subbed, I'd like it if they got well and truly booted. I had both sides of a partnership being problems at one point, and each was refusing subs for the other.

If you are director and you sub, the partner of the player to be subbed is not asked by the software if he likes his new partner. Only the player who is invited to sub is asked if he likes to, provided the subbing method is used where no names need to be typed in.

Quote

6. Besides accept and decline director calls, could I get a "hold on a moment" option that will both send that message to the person and make the director call go away only to come back in one minute? It's really annoying when I'm trying to fix somebody only to have to tell the people that also need help that I'm not ignoring them on purpose.

Better solutions for this problem ("List of calls") have already been suggested here.

Karl

[Gerardo]

jtfanclub, on Aug 18 2004, 02:59 AM, said:

1.  When I hit the button to kick people out of the queue before the tourney starts who aren't online, it would be really appreciated if it would also kick out people who are currently playing in other tourneys (usually because they subbed in).  I'm tired of having 36 pairs signed up and getting somewhere between 14 and 17 tables.  People who are offline might come back.  People who are playing in another tourney aren't.

Good idea

jtfanclub, on Aug 18 2004, 02:59 AM, said:

2.  I really really want to be able to set the minutes per board with a decimal points.  For example, my tourneys have 6 minutes per board, 4 boards per round.  When time is running out, usually about a third of the people are still playing the last board, which is no problem...I'm happy to give them a couple of extra minutes to finish.  But there's also almost always one or two tables still playing the next to last board, and I don't want them starting the last board and delaying the entire tourney for 6-8 minutes.  Unfortunately, if on minute #23 I change the tourney from 6 minutes per board to 7, and then on minute #24 somebody finishes the next to last board, it lets them start a new board.  If I could change it to 6.5 minutes per board, it wouldn't allow people to start new boards while giving the people on the last board an extra two minutes to finish.

That should be solved changing the meaning from minutes per board to minutes per round, IMO.

jtfanclub, on Aug 18 2004, 02:59 AM, said:

5.  When I asked for somebody to be subbed, I'd like it if they got well and truly booted.  I had both sides of a partnership being problems at one point, and each was refusing subs for the other.

I think they were exercising their right to try to finish the tourney with his/her p. They should get A- if that causes them to don't finish boards, though.

[jtfanclub]

Quote

That should be solved changing the meaning from minutes per board to minutes per round, IMO.

Good point.

Quote

I think they were exercising their right to try to finish the tourney with his/her p. They should get A- if that causes them to don't finish boards, though.

I'd never kick out a partnership for being slow. I was kicking them out for...other reason.

I could swear that as director I've seen "They" refer to N/S sometimes. I'll have to watch for it.

[jtfanclub]

Quote

Another argument for few board per round is that the annoying breaks between the round for fast tables are longer if you have more boards per round.

I'm using 6 minutes per board, with up to 4 minutes 'slack' to finish the last board. Slow players end up playing 3 boards per round and get an average on the fourth board. Fast tables end up playing 4 boards per round.

Let's say, for example, that a fast table takes 5 minutes per board and a slow table takes 9. In my tournaments, the slow table finishes 3 boards just in time, and the fast table ends up waiting 7 minutes after playing 4 boards. If we were playing two boards a round and I adjusted to the slow table, the fast table would have to wait 8 minutes per round, with twice as many rounds. If I adjusted to the fast table, I either have to adjust every other board by the slow table, which means its based on my judgment instead of their play, or I have to let them get the Ave-.

3.5 boards per round is either considerably faster than or much fairer than a two board a round set. And no, rushing the slow players in an effort to make it 'fair' isn't fair either.

Here, I'm using mine as an example. Let's use a much more basic set: 16 boards, 7 minutes per board.

With 2X8, the faster tables end up waiting 4 minutes per round. Assuming that they don't wait around after the last round, they end up waiting 28 (7x4) minutes. With 4X4, assuming they don't wait around after the last board, they end up waiting 24 minutes (3x8). So there's less waiting around with more boards per round.

With 2x8, you end up adjusting after each round, so with a slow table you end up adjusting 8 boards. With 4x4, you end up adjusting after each round, so with a slow table you end up adjusting 4 boards.

Therefore, there is less waiting, and less adjusting (which is far more unfair than playing only four other pairs, IMHO) by going with more boards per round.

Is there anybody running a 16 board match with 2 boards a round on a frequent basis? Or have people simply given up on that concept?

Quote

But if you have more tables, how do you define winning and losing? E.g. with 4 tables and 4 different results, what do you score for the middle results? I cannot see how this can work for more tables.

You use IMPs. The total IMPs for a table for a round is 0. When you're done with the round, the winning pair at the table gets a 1, the loser gets a 0. In tournaments, most end up using a 20 or 30 point victory scale instead of 0/1, but the effect is very similar because it takes only a few IMPs difference to have a drastic difference in the VP scale.

Quote

If you are director and you sub, the partner of the player to be subbed is not asked by the software if he likes his new partner. Only the player who is invited to sub is asked if he likes to, provided the subbing method is used where no names need to be typed in.

I wasn't aware of that. Thank you.

Quote

Better solutions for this problem ("List of calls") have already been suggested here.

Unless you're speaking of something I missed, list of calls helps the director prioritize calls. It doesn't let the players know that their call has been received but the director is too busy to respond at the moment.

[Gerardo]

jtfanclub, on Aug 18 2004, 04:27 PM, said:

Unless you're speaking of something I missed, list of calls helps the director prioritize calls. It doesn't let the players know that their call has been received but the director is too busy to respond at the moment.

People get "Director xxxxx has been notified" or something to that effect, when they make a call.

[mink]

jtfanclub,

abouth the waiting: the sum of all waiting is not as important as the length of a single break between rounds. If I have to wait 7x3 minutes, I do not care too much - probably I like to have a look at the 2 boads anyway. But if I have to wait 3x6 minutes this is rather long each time. And - I usually wait at the end of the tourney too - in order to see the result.

About your scoring system: I did not understand it! "You use IMPs" is not an explanation. This is like saying "you use numbers" when explaining how to multiply.

Karl

[jtfanclub]

mink, on Aug 19 2004, 05:16 PM, said:

About your scoring system: I did not understand it! "You use IMPs" is not an explanation. This is like saying "you use numbers" when explaining how to multiply.

Karl

Each of boards in the round are scored via IMPs.

Suppose, for example, that I have a total of +36 IMPs for the round, and you have -36. Then I get one point for the round, and you get 0, same as if it were +1/=1 or +100/-100. If there are six rounds, then at the end of the tournament we'll each have between 0 and 6 points.

This is, in effect a 1 point VP scale. Here's the 30 point VP scale:
IMPs Victory Points
0..... 15-15
1..... 18-12
2..... 19-11
3..... 20-10
4..... 21-9
5-6... 22-8
7-8... 23-7
9-10.. 24-6
11-13. 25-5
14-16. 26-4
17-19. 27-3
20-23. 28-2
24-27. 29-1
28+.... 30-0

So, if you win the round by 10 IMPs, you get 24 VPs. If you lose by 10 IMPs, you get 6. At the end of the game, the team with the highest number of VPs wins.

The advantage of VPs is that it keeps you from simply beating up on the little guys. Take two teams:

Team 1 wins the first match by 10 IMPs, goes on to face the winning half of the bracket, and wins by 10 IMPs.
Team 2 wins the first match by 50 IMPs, goes on to face the winning half of the bracket, and loses by 10 IMPs.

Using a pure IMP scoring, team 1 is +20 IMPs and team 2 is +40 IMPs, so team 2 is well on his way to winning. Using a VP score, team 1 has 24+24=48 VPs, and team 2 has 30+6=36 VPs. Team 1 is winning, although not by a huge margin.

In an open tournament like ours tend to be, you can often win or lose simply by who you get in the first round. If you're a top pair who faces a decent pair, you'll barely beat them (on the average) and face stiff competition for the rest of the tourney. On the other hand, if you're facing a pair who just learned the day before that aces take kings you can build up a huge lead which the rest of the field may not be able to catch, and the better players will never get a chance to play the newbiews if winners play winners.

Victory Point scales are very common...if you'd like, I'll find you more detail on them.

[mpefritz]

I think this system will lead to wildly unreliable results. Say a pair misdefends a game contract on the first board. That pair may then take wild risks to regain lost IMPs in that round in order to win the 1 VP. This will likely happen every single round because there is likely to be some abnormal result somewhere in the tournament on the first board of each round.

Now what you get is more and more shooting -- which can seroiusly change the scores at the other tables through no fault of the players at the non-shooting tables. yuck!

fritz

[mink]

jtfanclub,

now I understand your idea, thanks for explaining. I assume you gave that 30-victory-point-scale only to make your 1-victory-point-system clearer.

In your tourneys you have 4 rounds with 4 boards each. I think it would be likely there that you have several pairs with 4VP at the end, and therefore several winners. Maybe that will help to make the system popular.

But I agree with mpefritz about the merits of this scoring system: Pairs with a good score in the first board never bid a slam in the second if there is any risk, because they do not really need the additional points for the slam. And pairs with a bad score will take any risk. This affects the IMP scoring of the second board in an unforeseeable way.

And something else: imagine all board in a round are completely flat. Still, at some tables strange results will be produced, so that the IMPs for the majority of the pairs will not be 0 but slightly above or below. Who is above and who is below is pure luck then. And this luck decides about wether you get your victory point for the round or not. If you do not get it, you cannot win the tourney anymore, no matter how good or bad you perform in the other rounds. This sounds most unfair to me.

Karl

[hrothgar]

mink, on Aug 18 2004, 06:03 PM, said:

hi jtfanclub,

Quote

2. I really really want to be able to set the minutes per board with a decimal points. For example, my tourneys have 6 minutes per board, 4 boards per round. When time is running out, usually about a third of the people are still playing the last board, which is no problem...I'm happy to give them a couple of extra minutes to finish. But there's also almost always one or two tables still playing the next to last board, and I don't want them starting the last board and delaying the entire tourney for 6-8 minutes. Unfortunately, if on minute #23 I change the tourney from 6 minutes per board to 7, and then on minute #24 somebody finishes the next to last board, it lets them start a new board. If I could change it to 6.5 minutes per board, it wouldn't allow people to start new boards while giving the people on the last board an extra two minutes to finish.

This is a problem only if you have rounds of 3 or 4 boards. I do not like such tourneys, and I tell you why: In order to have a tourney that it as fair as possible the average skills of the opps of any given pair should be as close as possible to the average skills of all pairs of the tourney. The odds to achieve this increase if you increase the number of opps for a pair. This is done by decreasing the number of boards per round. So 1 board per round tends to be as fair as possible to this respect.

The statement is factually incorrect.

Tournaments can (and should) be modelled as a statistics problem.

A tournament consists of a set of of partnerships. Each partnership can be assigned a rating describing its strength. This population statistic can be used to establish a rank ordering of pairs.

During the tournament, pairs of partnerships are randomly matched. Each time that a pair of partnerships compete over a board can be treated as an indepent sample. The result of this event will in part depend on the relative strength of the two partnerships, however, there is also a random component. There are boards that are going to be flat regardless of the relative skill of the two partnerships. There will even be boards that favor weaker partnerships.

The resulting sample statistic can then be used to establish a rank ordering or the partnerships. Ideally, you want the sample statistics to approximate the population statistics as closely as possible. However, to do so you need to make sure that

(A) Individual partnerships content over a large number of baords
(B) A sufficiently large number of partnerships contest against each other.

The scheme that you recommend has pairs competing over a single board, which has the potential to significant bias results. I've never sat down and solved this problem since I never had good "objective" information regarding the standard deviation of partnership strength. Its also unclear what type of confidence interval would be considered to be appropriate. However, intuitively I doubt that a corner solution will prove to be optimal.

In any case, at the end of the day, there is going to be a relationship between

1. The number of pairs competing in a tournament
2. The length of the tournament, measured in boards
3. The statistical validity of the results

No way to dodge this bullet. My own preference is to hold relatively small tournaments that are able to achieve accurate results within sections. However, this doesn't necessarily allow players to purchase masterpoints as cheaply.

[jtfanclub]

Quote

now I understand your idea, thanks for explaining. I assume you gave that 30-victory-point-scale only to make your 1-victory-point-system clearer.

No, it's because a 1 point scale is easier to program. The big Swiss tourneys use a 30 point scale (which give virtually all the points to the winner unless it's very, very close).

Quote

And something else: imagine all board in a round are completely flat. Still, at some tables strange results will be produced, so that the IMPs for the majority of the pairs will not be 0 but slightly above or below.

Actually, that's worse in IMPs. If you didn't get to play one of the strange results, you can give up hope about halfway through the match.

Let's say, for example, that in boards one and two all N/S get -100 for each except one pair gets 1430 each time due to astoundingly bad leads. That one table, one pair has +32 IMPs, whereas the other tables everybody has about +/- 2 IMP. With victory point scoring, all the N/S players need to do is get 2 IMP better than their opponents over the next 2 boards, hardly an unmanageable feat. With IMPs, you may as well start the tourney over. h

[uday]

Hi JTFanclub

These are the ideas I will pursue

1. TD function to boot players in other tourneys as part of boot offline players
6. yes, TD call management needs much rework. Will fix. Don't know if i will include a :hold: option but i recognize the problem it is trying to address

These I won't

2. Decimal point for minutes per board. The code changes will ripple thru the server and I dont like the risk/reward/effort ratios.

5. We used to boot out people who were subbed out but i think the current return-to-lobby is better. Dont follow "each was refusing subs". As a TD you have 2 ways of subbing for a player A who is 9say) stuck.

Either type in REPLACE A WITH SUB in the edit screen, or
REPLACE A WITH (blank) . The former replaces A immediately whereas the latter asks for a sub from the sub list.




And these..?

3. Directions. There are a few cases where directions are needed. Most of the time, they are not, as Mink pointed out. Currently we're having trouble finding a good place to draw the directional. EMail fred@ with suggestions :) Should I change the History format to something other than SWNE ?

4. Victory points. New scoring formats are non-trivial. I dont see any demand for this except from you?


ui

[jtfanclub]

Sorry, Mink corrected me on #5. That will work fine, I just wasn't aware the two forms of subbing worked differently.

For things like VPs and decimal points per board, I have no idea what it will take in software, and no idea if other people want it or not. So I put it out there, and if it the demand/difficulty ratio isn't high enough, that's fine. Ultimatums these aren't.

For directions, the time I need it most is when I'm adjusting boards. Let's say that Person X slow-played board 3, and so I need to give an A+ to X's opponents on a never started board 4. I can't give an A+ to a pair by name, I have to give it to N/S or E/W. And short of actually calling up the empty board and looking at it, I don't know of a good way to do that: I keep getting the SWNE wrong on the previous screen (table/show history). And bringing up the hand record isn't fun, because when you get director calls during it it causes all sorts of messes.

I think if it were NS || EW I'd have a much easier time, especially with the || in there. If I confuse N with S, the worst is that people might look at it a little funny. If I confuse N with W, bad things happen.

Oh, and I have a #7, which should be pretty easy if I remember my Windows C++ days....

Instead of having the director calls pop up in the middle of the screen, could you have them pop up in the upper left? That way it'll take a lot more of them to cover what I need to see.

Thanks for your reply. I appreciate it. I'll email Fred in a day or two about directions and this last one.

[mink]

Quote

Uday: Should I change the History format to something other than SWNE ?

The current format of the bidding history is WNES, and I like this. All players are used to it. Please do not change.

Karl

[spwdo]

jtfanclub, on Aug 23 2004, 04:14 PM, said:

Sorry, Mink corrected me on #5. That will work fine, I just wasn't aware the two forms of subbing worked differently.

For things like VPs and decimal points per board, I have no idea what it will take in software, and no idea if other people want it or not. So I put it out there, and if it the demand/difficulty ratio isn't high enough, that's fine. Ultimatums these aren't.

For directions, the time I need it most is when I'm adjusting boards. Let's say that Person X slow-played board 3, and so I need to give an A+ to X's opponents on a never started board 4. I can't give an A+ to a pair by name, I have to give it to N/S or E/W. And short of actually calling up the empty board and looking at it, I don't know of a good way to do that: I keep getting the SWNE wrong on the previous screen (table/show history). And bringing up the hand record isn't fun, because when you get director calls during it it causes all sorts of messes.

I think if it were NS || EW I'd have a much easier time, especially with the || in there. If I confuse N with S, the worst is that people might look at it a little funny. If I confuse N with W, bad things happen.

Oh, and I have a #7, which should be pretty easy if I remember my Windows C++ days....

Instead of having the director calls pop up in the middle of the screen, could you have them pop up in the upper left? That way it'll take a lot more of them to cover what I need to see.

Thanks for your reply. I appreciate it. I'll email Fred in a day or two about directions and this last one.

hi,

When adjusting put in your preferences , player to the bottom is south, so when seeing a movie, or visiting a table u have your directions right, doesnt matter if u mix "E with W" or "N with S", it works just fine like it is.U need to see the movie i think before a adjust or u were at the actual table where u witnessed slowplay, either way if u have player to the bottom=south u wont have problems with directions.


Marc

[jtfanclub]

Quote

U need to see the movie i think before a adjust or u were at the actual table where u witnessed slowplay, either way if u have player to the bottom=south u wont have problems with directions.

When the board was never started? What's the point in that? If I know that player X's opponents didn't even show up until the middle of the 2nd board, I don't need to call up the handrecord to figure out which side needs adjusting- I just need to know who sat where. For that matter, if I'm watching a table to let them play to the end of the round, then I want to adjust it, the table is now gone and I need to go back to make sure I have which side was N/S right. I don't want to need to bring up the hand record- that part I figured out when they were playing.

Maybe most people don't have any trouble with telling which middle person is N/S and which is E/W. When I'm running a large tourney, and I have a half dozen adjustments to make in under a minute, I certainly do.

EDIT: If you go into the Hand Records for Tourneys- Main Bridge Club (Web), the format is:

Time: North South East West Result Points Score Movie Traveller

To me, all of this is useful and easy to see information. I would absoutely love it if the Tables/Show History had exactly this format.