BGonline.org Forums

New standard method for resolving 3-way ties on the ABT

Posted By: Rich Munitz
Date: Thursday, 21 June 2018, at 5:02 a.m.

In Response To: New standard method for resolving 3-way ties on the ABT (Colin Owen)

I'm not particularly hung up on terminology. Weird or unusual. Yes. It is not a conventional solution. We agree there. Not a big deal.

Where we seem to disagree is on the practicality. Both this approach and your Round Robin approach are exactly the same for the first two matches. Play match 1 and the winner of match 1 then faces the 3rd player. Identical to here. In our format, we're done 100% of the time, with a reasonably equitable resolution.

Round robin is done 50% of the time. Suggesting that they are often done at the same time, but there is the possibility it might take longer is an "unusual" way of saying that on average, the 3 players are not done with play for an additional 2 hours.

An 11 point match is expected to take about 2 hours. How many extra matches are expected from a round robin format over the exactly 2 matches in the described method? I now realize that my original statement of 4 matches (2 extra matches) is incorrect, because the 3rd RR match is not always needed. The corrected answer is 3.33.

But let's simplify and just look one RR ahead and assume it will be resolved by then without further repetition. Half the time it is done in 2 matches. If not, then half of the remaining time it is done in 3. The other 25% of the time there is a tie and we are back at square one with a 3-way tie and have to play another entire Round Robin. Suppose that Round Robin completes without another tie and we are done on average in 2.5 additional matches. So the expected number of matches here is (.5*2 + .25 * 3 + .25 * 3+2.5) = 3.125 (a bit short of the long-term 3.33 result). That's an expected 1.125 11 point matches above and beyond the 2 matches asserted to be what should "usually" happen. If you expect to play an extra 1.125 11 point matches you must expect to add another 2.25 hours to your tournament schedule. Sure, that expected 2.25 hours might turn out to be zero. But that 2.25 hours might also turn out to be 4.5 hours, or 6 hours, etc.

In summary, I was completely correct to say that a 3-way tie is expected take 2 hours longer to produce a result if you try to resolve it using a Round Robin approach. Round Robin is valid. It is theoretically a more perfect resolution. But why on Earth would anyone in practice want to pay that price in time for such a small incremental benefit? I sure wouldn't as a player, and no director should impose such a burden on their attendees.

The USBGF has taken the position that there is indeed a better way to resolve these situations and have adopted such a standard. To everyone else, it is simply a logical and practical suggestion.

Post Response

Subject:
Message: