Life In 19x19

A six round McMahon tournament was played. The scoring criteria were McMahon points, SOS, SOSOS.

A 2 dan player was put in the top group, but he wasn't playing all the rounds. He only played the first to games and lost both. Two players finished with 5/6 result at the top with equal SOS, so the trophy was decided on SOSOS. One of the two players had played this 2 dan on the first round and thus his SOS suffered.

I am thinking, whether this kind of mistake could be conpensated in future tournaments. There are several options to consider:

Do nothing.
Lower the 2 dan´s McMahon score retrospectively. (this would hurt the opponent even more)
Count the 2 dan´s opponents as having a bye win instead of winning the 2 dan.
Anything else?

What do you think?

edit: 2 dans -> 2 dan´s

It seems difficult to correct such a mistake without the risk of making a new one.

Which part was considered the mistake ?

Hello,
With the OpenGotha software, I use to grant 0.5 point of SOS for every round missed by an opponent, so that the SOS of the player gives an idea of the average score that the opponent could have got if he had played all the rounds.
In the meantime, the absent player has zero point for missed rounds.

The advantage are :
-The SOS is as right as it can be, granted the information we have
-Absent players are not rewarded if they miss a round

The drawback :
If a player misses some rounds, then plays again, he is paired, according to his McMahon score, with too weak opponents.
But it is also an advantage if we need to reward the top three players, since it avoids absent players messing with the top of the tournament, where we need the ranking to be accurate.

The european federation advises to register strong players who plan to miss some rounds one or two points below the top group.

EdLee wrote:Which part was considered the mistake ?

Letting a player, who was not going to play all rounds to start from the top group.

Pio2001 wrote:Hello,
With the OpenGotha software, I use to grant 0.5 point of SOS for every round missed by an opponent, so that the SOS of the player gives an idea of the average score that the opponent could have got if he had played all the rounds.
In the meantime, the absent player has zero point for missed rounds.

The advantage are :
-The SOS is as right as it can be, granted the information we have
-Absent players are not rewarded if they miss a round

The drawback :
If a player misses some rounds, then plays again, he is paired, according to his McMahon score, with too weak opponents.
But it is also an advantage if we need to reward the top three players, since it avoids absent players messing with the top of the tournament, where we need the ranking to be accurate.

I know this.

The european federation advises to register strong players who plan to miss some rounds one or two points below the top group.

The problem occured when this advice was not obeyed.

Javaness2 wrote:It seems difficult to correct such a mistake without the risk of making a new one.

That risk exists.

Let me try another option:
Calculate the average of McMahon points of the players who played in the top group excluding the 2 dan, who did not play all the rounds. Use this number (possibly rounded to nearest 0.5 points) as contribution for the opponent´s SOS.

If the rounding takes place, this usually reverts to a simpler option:
For the opponent of the first round count 0.5 wins for each subsequent round, not rounding to an integer. In this case it would be 2.5 wins.

Anyway, things get messy, if a plaeyr drops out from the tournaent and then re-enters.

Letting a player, who was not going to play all rounds to start from the top group.

Do you mean it was known at the time of registration that the player would not play all the rounds ? What if it wasn't known at the time of registration, but someone skips a round or more unexpectedly ? What if this isn't a McMahon system, and at the time of registration it's known someone will take a 'bye' ?

EdLee wrote:

Letting a player, who was not going to play all rounds to start from the top group.

Do you mean it was known at the time of registration that the player would not play all the rounds ?

At least the player knew it. I don't know whether it is the player's or organisers' fault that the player was entered in the top group.

What if it wasn't known at the time of registration, but someone skips a round or more unexpectedly ?

The EGF tournament rules were in effect. The player who breaks the rules may get a penalty.

What if this isn't a McMahon system, and at the time of registration it's known someone will take a 'bye' ?

The EGF tournament rules can also be used in other tournaments. If at the registration someone says he wants to take a bye, the organisers decide the action suitable for the tournament system.

Admit that your data ( SOS and SOSOS ) is corrupt, and therefore discard it.

That leaves you with no means of tie-breaking. So declare them joint winners, and pay each of them the full prize ( cash or trophy or whatever ).
This means that if there is a cash prize, you have to add the difference between 1st prize and 2nd prize from your pocket. If there are trophies, you have to pay for another one and deliver it late to one of the winners.

Joaz Banbeck wrote:Admit that your data ( SOS and SOSOS ) is corrupt, and therefore discard it.

That leaves you with no means of tie-breaking. So declare them joint winners, and pay each of them the full prize ( cash or trophy or whatever ).
This means that if there is a cash prize, you have to add the difference between 1st prize and 2nd prize from your pocket.

This was done.

If there are trophies, you have to pay for another one and deliver it late to one of the winners.

The trophy is a rotating one and goes from the previous winner to the next one. A duplicate was not bought.

I think the such issues are inherent in systems where stuff like McMahon and SOS are used for determining final sequence of participants. It caters to our need to be able to say stuff like "I scored higher than you" even if "we never played against each other and we both have the same score." And it is INHERENTLY inaccurate and often prone to completely random events or even arbitrary.

Personally, I see nothing wrong with just having shared places. Just look at straight-up score and assign places. Share prizes if necessary.

If there is a need to more strictly determine the sequence of players - for qualification purposes, for example - let the players in question play a game to decide who wins the spot.

I know this is not a popular opinion, and not always that easy (and has little chance of being accepted) - but this is what I think is the lesser evil, in this case. It might put additional strain on the organizer (potential need for extra games) - but I think this is (more) fair to the players.

PS>
Considering the above, the main mistake the organizers made was running a tournament in which McMahon, SOS and SOSOS were considered as the means to handle tiebreaks.

Just my opinion.

Bantari wrote:Considering the above, the main mistake the organizers made was running a tournament in which McMahon, SOS and SOSOS were considered as the means to handle tiebreaks.

There is nothing wrong in using SOS and SOSOS as tie-breakers, as long as they are correctly defined and calculated.

Saying that the SOS of a player is "the sum of his opponents' scores" is as wrong as saying that his points on the board are "the number of intersections in his territory".

For the SOS of a player, I think I can give a better definition :
It is the sum of his opponent's scores, plus the number of missed round times his initial score, plus half the number of rounds missed by his opponents.
It is not a correct definition, though, because it doesn't take into account the points scored by bye players, that should not be counted in the SOS of their opponents. And it relies on the fact that a player scores zero points when he is absent (mind that an absent player (willingly not playing) is not the same thing as a bye player (unwillingly not playing)).

But I don't know how to define the SOSOS.

Pio2001 wrote:There is nothing wrong in using SOS and SOSOS as tie-breakers, as long as they are correctly defined and calculated.

A player can play perfectly and lose out on first place due to circumstances beyond her control. That's the main thing I find wrong with them.

dfan wrote:

Pio2001 wrote:There is nothing wrong in using SOS and SOSOS as tie-breakers, as long as they are correctly defined and calculated.

A player can play perfectly and lose out on first place due to circumstances beyond her control. That's the main thing I find wrong with them.

MMS and SOS are measuring similar things, that's why SOS is a commonly used tiebreaker. I think everyone understands that tiebreakers should only be used when they really need to be used. Generally, it is better to award people with the same MMS the same prize money, the same qualifier points. Sometimes you cannot, and you have to break ties. In that situation, using SOS is about as good as you can get, particularly in a McMahon tournament with a large number of rounds.

We seem to have slid off topic though