Tiebreakers

I’m currently running a scholastic tournament at one of my school clubs. There are 16 players with no ratings at all, so for the first round I simply ordered the players alphabetically and took it from there. After the first round I’ve been pairing by score groups.

I’m not using pairing software, since it’s a small tourney and I’ve done this before with rated tournaments. There will be five rounds, and the players are 5th and 6th grade students.

My question is, in this situation, what’s the best way to break ties at the end when it’s time to pass out the awards?

I’ve heard I can use the cumulative method, since it’s simple and fast, but I need a second tiebreaker if that doesn’t do it. I realize the modified median system should be used first, but given my time constraints, I need to be able to break the ties within a 10-minute time frame. Besides, given the explanation of this system in the rulebook, I think I’d be better off getting out my copy of the Harkenss book for clarification!

Looking forward to hearing from you!

Radishes

If you’re going 4 rounds with the 16 players, I’d recommend the Solkoff, which is a variation of the Median. The Solkoff doesn’t throw out any games (extremes) like the Median does. Simply add the final scores of the opponents of the tied players. The Median isn’t recommended for tournaments of less than 6 rounds.

Yeh, cumulative is the way to go if time is of the essence.

You can even calculate cumulative tie-breaks before the last-round games finish. Just calculate it based on all rounds EXCEPT the last. This works because:

For two players with the same score at the end of the tournament, if player A has better cumulative tie-breaks than player B, then, and only then, player A will also have better tie-breaks than player B if the last round results are omitted.

Bill Smythe

Just a word about ranking unrated players:

Alphabetically, or randomly, is fine for the first 2 rounds. After that, you can rank the unrateds in order by cumulative tie-break so far. For example, in the 1-point group after 2 rounds, you can consider players with a win in round 1 to be stronger than those with a win in round 2.

Bill Smythe

Okay, I had wondered about ranking them after a few rounds. But I guess I don’t follow your logic here. As I understand it, the cumulative of both players you talk about would be 1, wouldn’t it? 1+0=1 for one player, and 0+1=1 for the other is the way I read it for doing it the cumulative way, unless you’re talking about using the scores of the opponents instead of their own scores.

Radishes:

A win in the first round and a win in the second round is not the same cummulative score. A player who wins the first game and loses the second has a cummulative of 2.0, that is, after round one he has one point and after round two he has one point, so the cummulative is two. To go a step further, think of a player winning his first four games. His cummulative is 1+2+3+4 =10. It does make a difference and the tiebreak makes sense. Take two players who after four rounds have two points each. One won his first two games and the other won his last two games. The first one has a cummulative of 1+2+2+2=7 while the other one has a cummulative of 0+0+1+2=3. Why such big difference? In round two the first player probably played someone who was 1-0 while the second played someone who was 0-1. In round three the first player probably played someone who was 2-0 while the second player played someone who was 0-2. In the fourth round, the first played a 2-1 and the second player played a 1-2. As you can see, the first player has had a much tougher road to the same score, therefore deserving of a higher tiebreak (and is probably a better player)!

'Hope this helps to explain…

I did check that out, and understand a bit better now. I had thought that was the way it worked, but I needed some clearer explanations.

And I knew you folks could straighten me out!

Radishes

Personally, if you only have two players tied, I would check to see if they played head to head already first. That result is a good tiebreaker. I don’t tend to like any tiebreak system. That is why I give cash prizes.

Head-to-head can be counterproductive.

If player A defeated player B, then player B may have otherwise had a perfect score, while player A lost to some other clown not even involved in the tie.

Bill Smythe

Head-to-head tiebreaks have two major advantages:

  1. Parents understand how they work (and feel they are fair)
  2. They don’t require a PhD to explain how they work

There is never any good tiebreaking system, as its’ the first time for the player and the parents (majority of parents and players), at a in school tournament. Having a bad tiebreak, or the feeling of a bad tiebreaking with the parents. It will be the first and last time they ever come to a scholastic tournament. How the tournament is organized, how the awards are given out: becomes a budget question on the short fall of trophies and awards needing a tiebreaking system.

I wonder if I can come up with a non-standard system. Something using both the results and the opponent rating.

Look at the result of the game if the opponent is higher rated multiply it by 2. If lower rated, multiply it by 1. Sum the totals.

So something like this against our imaginary 1350 player:

Rating Result Multiple Subtotal
1500 0 (Lost) 2 0
1200 1 (win) 1 1
1450 1 (win) 2 2
1300 0.5 (Draw) 1 0.5
4-Round Result 3.5

Thoughts? I don’t recall this being a standard system but it seems somewhat more fair.