Improving the tournament process

[TylerE]

I have a question: is the draw of ACBL tournaments really the masterpoints?

Sadly, yes.

[TimG]

Another question: could someone besides BBO get a tournament on BBO sanctioned by ACBL? That is: does BBO have exclusive rights to ACBL games run on BBO?

[uday]

BBO have exclusive rights to ACBL games run on BBO?

 

I'm sure Memphis would not mind in the least if BBO allowed another ACBL-sanctioned club to run on BBO.

 

But it would be a hard sell for me...

 

U

[awm]

I have a question: is the draw of ACBL tournaments really the masterpoints?

 

Why hasn't someone else set up the kinds of tournaments people are suggesting here? Actually, I'm sure there have been tournaments such as these run, but why hasn't anything caught on and taken off?

I'm sure the masterpoints are part of it. But there are many free tourneys which get hundreds of pairs signed up -- it's not like the ACBL tourneys are crowding the others out.

 

In addition to the masterpoints, I'd say there are a few nice things about ACBL tourneys:

 

(1) They have a fairly professional director. In contrast, free tourneys sometimes have playing directors or one director for several hundred tables (but you get what you pay for) and the other pay tourneys can be hit or miss. This is not to say that the directors always make great rulings, but at least it's a good bet that the director will come to the table when called and that there won't be obvious favoritism in the ruling.

 

(2) They are run at predictable, regular times.

 

(3) They're run on a fairly strict clock, meaning everyone finishes at the same time, you get results in a timely manner after you're done, etc.

[pigpenz]

In fact it's interesting, it seems like sometimes the ACBL tourneys play 6 rounds of 2 and sometimes 4 rounds of 3. It appears to depend on the total number of competing pairs... but there are always way more than 12 pairs so it's not like we couldn't always play 6 rounds of 2... it has something to do with section sizes etc.

the more rounds the longer it takes to play so it cant be fit into an hour schedule

due to director calls and sometimes needing to get subs etc...so the 4x3 maximizes the chances that everything is done right before the hour so that the next game starts

[hrothgar]

BBO have exclusive rights to ACBL games run on BBO?

 

I'm sure Memphis would not mind in the least if BBO allowed another ACBL-sanctioned club to run on BBO.

 

But it would be a hard sell for me...

 

U

And this is just why I argued LONG LONG ago that it was a mistake for BBO to integrate vertically and start running tournaments.

 

From the sounds of things, you're running into conflict of interest problems because a third party offering ACBL sanctioned tournaments might threaten your monopoly position...

 

In all seriousness, can you legitimately explain how blocking third parties from offering ACBL sanctioned tournaments improves the service that you are offering your customers?

 

More importantly, you really might want to consider that customer service is how you compete in these types of markets. They day that you forget this is the day that some new service is going to be able to establish itself as a viable alternative.

[0 carbon]

I agree with 1-board rounds. Yes, more delays - half the time a delay every board!

 

But can't BBO be flexible with the movement? IOW, instead of having a fixed movement with everything planned from the start, match pairs together when they finish. Then the only constraints might come toward the end when you have fewer unplayed pairs to play against. Not a problem in big tournaments. You would also have to watch ahead so that you preserved at least one unplayed pair for everyone for every round.

 

After all, there are no physical considerations (Boards up, players down, duplicate decks, ....) to worry about.

 

I also support having the choice of "Swiss" style Victory Point pairs where pairs of each round are matched on scores. This, of course, means guaranteed delays waiting for the final result, but why not try it, clocked? A fair scoring system would have to be worked out, though....

 

tOM

[uday]

match pairs together when they finish.

 

That's what the "unclocked" movement already does. Playbacks are routine in an unclocked movement and some pairs will end the T long before others.

[awm]

Last night I played in a free IMP pairs tourney with the "swiss with rematches" movement.

 

I thought it was a good illustration of why I don't like this movement at all, with us being on the "screwed over" end (obviously I have been on the other side of this a few times before as well).

 

After racking up about 8 imps in a fairly flat first round, we clobbered our second round opponents to the tune of 15 imps. At this point, we moved up to "table one" and played against what was probably the best pair (other than us) in the tourney. When playing against a good pair, they are usually not going to roll over and give you a zillion imps. And they will occasionally win imps on a hand by doing something clever (i.e. avoiding a bad slam that the field bids) where you can't really do much about it.

 

Okay, so we're at table one. We play the best opposition, we beat them by a couple imps. Now we move for the fourth and last round and guess what -- we are still leading, still at table one, and playing the same pair again. Three more hands at the top table, we beat the other pair by a couple imps for the second time in a row.

 

Now the tourney is over. We finish... second. We have been passed by the pair we clobbered in the second round! If you look at "strength of opposition" we have faced much tougher opponents since they were down near the bottom after giving us 15 imps and working their way back up. Meanwhile, we were paired twice in a row against a strong pair at the top table, struggling for every imp. The pair who had the misfortune to play us was probably even more "screwed" by this than we were, dropping to fourth in the standings.

 

I'm really not convinced that this movement is better than the semi-random pairings we see in the ACBL games.

[qwery_hi]

Last night I played in a free IMP pairs tourney with the "swiss with rematches" movement.

 

I thought it was a good illustration of why I don't like this movement at all, with us being on the "screwed over" end (obviously I have been on the other side of this a few times before as well).

I guess you probably had the most fun of them all playing against top opponents - Just saying

[helene_t]

If you look at "strength of opposition" we have faced much tougher opponents since they were down near the bottom after giving us 15 imps and working their way back up.

Right. The good news is that the other pairs in the top end will also face strong opposition. It is "unfair" in terms of comparing the strong pairs to the weak ones but nobody cares about this. In a random movement, you can in principle have the misfortune of playing against strong opposition all the time, while others who play at approximately your level will face weak opposition.

 

Ideally, if the movement isn't perfectly balanced (a tournament with much more pairs than rounds cannot be anything near perfectly balanced), something like the Lehman rating rather than gross IMPs should be used for the ranking.

 

It may be interesting to do some simulations to shed light on this issue, but my gut feelings are

- swiss is better than random

- swiss is slightly better than rematch-swiss

- for tournaments with a large pairs/round ratio, using a more rational evaluation than gross IMPs could potentially gain more in terms of fairness than the swiss format can (but of course, one does not exclude the other).

 

But I could be very wrong about the above.

[qwery_hi]

Swiss can be gamed too. When I played in chess tournaments, some people used a strategy of losing in round 2 in a 9 round event so they could easily win rounds 3 thru 6.

[uday]

The scenario described ( 3rd place leapfrogs into 1st while 1st/2nd are beating each other up ) could easily happen in real life Swiss Teams, couldn't it ?

 

I suppose one could argue that +3 imps at Table 1 is harder to achieve than +3 imps at table 10 ....... but is that true for the pairs at table 10 ?

 

It almost feels like imps won should be multiplied by something, so that you get a bonus for winning the same # imps at table 1, slightly less at table 2, etc .

 

u

[hrothgar]

The scenario described ( 3rd place leapfrogs into 1st while 1st/2nd are beating each other up ) could easily happen in real life Swiss Teams, couldn't it ?

 

I suppose one could argue that +3 imps at Table 1 is harder to achieve than +3 imps at table 10 ....... but is that true for the pairs at table 10 ?

 

It almost feels like imps won should be multiplied by something, so that you get a bonus for winning the same # imps at table 1, slightly less at table 2, etc .

 

u

Once again, I will reocmmend the work that Gerben did modeling Swiss Team for the Aussies.

 

He did some nice work modelling just these types of adjustment factors.

[awm]

The scenario described ( 3rd place leapfrogs into 1st while 1st/2nd are beating each other up ) could easily happen in real life Swiss Teams, couldn't it ?

 

I suppose one could argue that +3 imps at Table 1 is harder to achieve than +3 imps at table 10 ....... but is that true for the pairs at table 10 ?

 

It almost feels like imps won should be multiplied by something, so that you get a bonus for winning the same # imps at table 1, slightly less at table 2, etc .

 

u

Of course it could. I am not complaining about the BBO software directly here, or about the directing or anything like this. I am trying to counter the view that swiss is somehow a "more fair" format and that ACBL tourneys (which currently divide people in sections fairly randomly and then just move through the section) should switch to a swiss format.

 

I do think that having rematches made this worse -- if there are two pairs who are much better than everyone else, they will be forced to play each other over and over until someone catches them, which seems kind of a weird way to do things. To look at it another way, in a fair tournament format you should always want to have the highest score possible going into the last round. Suppose I tell you that in a "swiss with repeats" the top four teams have 25, 25, 23, and 10 IMPs total and that the IMP scores are roughly reflective of skill level. Which team do you want to be? Which team has the best chance to win the whole thing? Odds are it's the third place team with 23 IMPs, and I think there's something wrong with that!