Solkoff, Modified Median, Median and Sonnenborn-Berger all depend on how well your opponents do, which is out of your control and thus a bit unfair (why should you be penalized for something you can’t control?). The closest to having your opponent’s results in your control may be what happened in 1991 when the IL state K-8 champion had better tie-breaks (with a 6.5-0.5 score) that the other player, and did so partially because of the post-game discussions with the beaten opponents that was geared towards keeping them from becoming demoralized and playing poorly in the rest of the tournament. Whether or not it actually made a difference is debatable, but it was the only time in the year that the tie-breaks went in her favor in the tournaments for which she tied for first.
Cumulative used to be the primary tie-break (easy to calculate) and got bumped down once people realized it often depended on just how lucky you were in pairings (thus the unfairness).
A blitz play-off (now favored for breaking ties between perfect scores) measures a different skill from playing in a slow time control.
Kashdan favors winning against a weak opponent and losing to a strong opponent as being preferable to two draws against equal or stronger opponents.
More blacks is due entirely to how the pairings worked out.
A coin flip is entirely out of the players’ control.
Note however that, even though they are unfair (due to depending on things you can’t control or measuring different skills from the main event), they are still blindly unbiased.
But that’s not the definition of fairness. If I roll a 6 sided die - the number that comes up is out of my control. But if the probability of each event happening are the same - then the die is FAIR.
Fairness is RELATIVE, not ABSOLUTE. This can confuse people because they inadvertently replace fairness with apparent consistency (i.e. they treat everyone the same). But consistency is not necessarily fair because the circumstances might be slightly (but significantly) different from case to case. (One can treat fairness as absolute by saying "All other things being equal…)
So, going back to the die example, maybe we have a die where the odds are unbalanced for each side; the die is UNFAIR - but if we change the way that the game is played so that the reward is different for each face - we can still make the game FAIR. In this case we have acted INCONSISTENTLY to restore FAIRNESS overall.
So again Jeff, I don’t see the tie-break systems as implicitly unfair for any player going into a tournament. I DO see that the systems are imperfect, such that as information changes they can provide apparently contradictory results because they provide different ways of measuring performance to break ties.
A nice story, but not related to the implicit fairness of the tie-break systems. What its related to is the importance of opponent performance FOR ANYONE involved in a tie. That’s fairness.
Fairness has nothing to do with control or the skill measured. Fairness has to do with treating the same circumstances in a consistent way so that one or another person is not awarded an unjustifiable advantage/disadvantage, or in balancing out unjustifiable circumstances.
We could create a tie-break means that measures a very unrelated skill or no skill at all. For example, ALL ties could be decided by lot. That would be FAIR, but it would also be UNSATISFYING and potentially UNWISE because the point of a tournament is to decide place by MERIT and breaking ties by lot doesn’t follow that tournament principle.
The situations you describe relate to the IMPERFECTION of tie-break systems, but they are applied uniformly and consistently and appear to be FAIR.
Using the wrong term with parents can be a truly bad idea. Why would we knowingly tell anyone that we have purposefully chosen a system that is UNFAIR? That would simply make us either fools or cheaters, Jeff.
It’s something different to say that unfortunately all systems have arguable imperfections - so that the best we can do is to choose systems that seem to have the least imperfection.
Given the advent of computers - I’m surprised that no one has put in tournament software a way to combine tie-breaks rather than sequence them. For example, following score, we could calculate each person’s final position based on each (selected) preliminary tie-break, and then take a weighted-average of those final positions as the actual tie-break. In theory, this is taking a weighted agreement among the tie-breaks for who is first, second etc. and is therefore arguably better than a sequential system, as it washes out imperfections and focuses on place agreement. The weighting would allow us to argue that certain systems are “more imperfect” and should therefore carry less weight.
I am using unfair in the most commonly perceived form (out of the control of a person with potentially adverse effects - similar to unjust when a person can be penalized for something the person had no control over).
Per your definition, if Bob is driving on a busy city street and a car in the opposite lane swerves across the median into a head-on collision, the result is fair because it could have happened to anybody driving in that lane and by chance it happened to Bob. I’d disagree that fair is a word that Bob or his family would consider appropriate.
Also, your die example is actually pretty good. The difference is that in a chess tournament each player has a different set of opponents, kind of like different dice, and each die has a different set of odds for the possible numbers. So my set of dice might be predisposed to roll higher total numbers than your set of dice. If the dice are unfair (your term when used for uneven results of rolls) and the tie-break system does not take that unfairness into account then the tie-breaks are perceived as unfair. They are, however, blindly unbiased because the unfairness is not targetted at specific people.
First, I disagree that you’re using it in the most commonly perceived form, although I agree that you’re using it in a commonly mis-informed way. I think FAIRNESS is commonly misunderstood and misused.
You’re also applying it in a situation that makes no sense. (The player is aware of the circumstances when they enter.)
No one is being penalized. Your analysis is adding a moral statement to the calculation of tie-breaks that doesn’t exist. In actuality, it’s fair, because all things being equal, everyone is being treated the same way. You are horribly mis-characterizing what is actually happening, and that’s a dis-service to all players and TD’s and I’d ask you to change that.
Players “pay their money and take their chances”. If you put money into the stock market and the stock goes down, was that “unfair”? OF COURSE NOT.
Terming actually possibilities that we knew all along as unfair is nothing more than pandering to selfish attitudes. Help people to understand the reality and the big picture, and it works better long-term for everyone. Why go around promoting that Tie-Break systems are unfair? Frankly, I view such an approach as very poor judgement.
Barring certain circumstances, its completely fair. It’s very unfortunate. It FEELS unfair. But it is not ACTUALLY unfair. (The certain circumstances are whether or not there were any legal violations occurring that contributed to the accident.)
Seeing that I’ve actually experienced those type of circumstances, I agree with how they feel. But I don’t make the mistake of assuming that’s how they ARE. I lost a brother in such an accident. It felt very unfair, but what happened, happened. And there was no fault (other than perhaps his own.)
I’ve also known people impacted by murder. That WAS unfair, and someone was punished as a result.
There is a difference between something FEELING unfair, and being ACTUALLY unfair. You enter a tournament, you know the tie-breaks just like everyone else in the tournament. Their application is fair. The result is fair.
First, seeing that the prizes and ratings weight constructs (such as prize distribution and pairings) based on the uneven strengths of the players, but that (in theory) all players know that construct upon entering, at a macro level the tournament functions in a fair way.
Moreover, Jeff, at the time of entry generally none (or at least an insufficient amount) of the information is known. (Even if it were, players are limited in their possibility to act upon it, so their is no implicit unfairness or anti-selection.) Knowledge impacts probabilities. IF their is unfairness (which would be difficult to demonstrate at best) it can be known ONLY IN HINDSIGHT.
So no, the argument you make doesn’t apply.
(BTW consider the parable of the laborers. Matthew 20:1–16. Is the treatment of any laborer unfair?)
The characterization of Tie-Break systems as unfair is a poor choice of words to use with players. Tie-break systems are limited, they each have shortcomings. Consequently they sometimes create results that in HINDSIGHT may feel unfair. But they each were known in advance (or had the opportunity to be known) and they were applied consistently. Thus, they are fair.
First of all, I don’t give a flying flugelheimer whether tie-breaks are fair or unfair, consistent or consistent, etc, and the extended discussion on this unimportant side-issue is getting extremely annoying.
Getting back to the main point –
Aren’t we all overlooking something?
The crosstables on MSA, while there for the organizers, are there even more for the players.
If we give organizers too many options, every crosstable will appear TO THE PLAYERS to be different from every other. Some will be ordered by score, others by rating, others by tiebreak, others alphabetically, others who-knows-what.
A worthwhile long-term project might be to give each viewer (whether player or organizer) the option of determining the listing order for the crosstable he is viewing. After choosing the crosstable, the viewer could be asked, via radio buttons, a question along the following lines:
For the crosstable you have chosen, in what order do you want to see the players listed?* by score, then by rating for players with equal scores (the default)
in alphabetical order
by score, then alphabetically for players with equal scores
by score, then by tiebreaks as submitted by the organizer
in the standings order submitted by the organizer
The last two options would be included only when (and if) the day comes when organizers are able to submit the necessary data, and even then, would be grayed out if the organizer chose not to submit such data.
I’m “fairly” certain I’ve said that a few times, and have noted that the players want the option, which is why organizers end up posting results online (since MSA doesn’t do it.)
I don’t think I ever argued to make it an organizer option.
In one of my posts I gave a similar longer term argument, so I agree. Having a short-term, disclaimed, standard tie-break option in addition to the current default was only a suggested temporary step, and was noted as such in a few posts as that.
Have we run this discussion into the ground yet???
The only part of it that still is holding my attention and giving me a few laughs is the discussions about people contacting the organizers if they have any questions. If I’ve been hired as a TD, I’m expecting to be doing tie breaks, etc., and with a few exceptions the organizers who have hired me to run a tournament for them would have no idea how the tie-breaks were calculated for a specific tournament since they didn’t do it. I don’t want MY contact information on MSA for people to send complaints as to why tie-breaks worked the way it did.
Perhaps the easiest way to provide the final standings with tie break is to add a spot on the event header where a link to the organizers site where the standings in tie break order are shown. Perhaps something like the following in each section panel
Crosstable data is NOT SHOWN in tiebreak order and does not reflect
any distribution of trophies or prize monies. Visit [link] for standings in tie break order.
Crosstable data is shown as reported to the USCF by the Tournament Director.
Please contact the TD to report any crosstable errors.
The additional text would only appear in any section for which the TD/organizer provided a link. I don’t know what to do for links that disappear. Keeping this updated would take a lot of work. As I consider this, perhaps best to either continue as we now do it or add the tie break info.
Its an interesting point, Ernie. Personally, I still favor doing what we can to drive web traffic TO our site rather than FROM it, so long-term I’d rather see us have the ability to have that information in MSA.
I like the link idea. As for disappearing (or hijacked) links, the viewers themselves are the best defense. As soon as a broken link is reported (perhaps even here on this forum), somebody (either SysAdmin or the moderators or somebody in the office) could check it out and remove it.
If we add tiebreaks, it should be only whatever tiebreak info is supplied by the organizer. Otherwise, there is about a 99% chance that, somewhere along the line, there will be a discrepancy. And that would increase, not decrease, the problem (if there is a problem) of parents, coaches, etc questioning or complaining about the standings order.
In fact, if TDs who use pairing software also use the prize report feature built into the software, the prize report itself (which is basically a delimited text file, I believe) could just be uploaded as part of the MSA submission. Converting such files to HTML is actually trivial - and it might not even have to be converted to display properly.
Yes, setting up a correct prize report can be a little involved for larger tournaments with lots of prizes. However, the TD who invests that time on the front end will experience two time savings on the back end.
Prizes are calculated automatically.
Prize reports - with tiebreaks - can be posted ASAP onsite (and even online, if the TD has a net connection). What are the benefits of this solution?
No added work for software developers - the functionality already exists. Minimal added work for TDs on the front end - and really, setting up a prize report is not hard at all, especially if you RTFM beforehand. Reduced work for TDs on the back end - all the prize information can be generated automatically at the end of a tournament, so awarding prizes and reporting prize-ordered results would be a snap. Minimal added work to reconfigure MSA - just add a field to the rating report form that allows uploading of prize reports, and arrange for them to be stored.Greatly reduced risk of user error - no worrying about entering custom tiebreak sort. Maximum information to MSA readers - they can find out exactly who won what prize, and what tiebreaks were used to calculate those prizes.
Now, there are three things that, IMO, should be noted with this proposal.
First, there would also need to be an optional field added to the rating report submission form that would allow an organizer to submit a link to an external site for display of results. Some events use non-standard tiebreaks that simply cannot be calculated by a prize report.
Second, in the benefits list above, I mentioned that there would be minimal work to reconfigure MSA and the rating report form. In this case, “minimal” does not mean “trivial”. But this is the least labor-intensive change, IMO, of all the ones proposed so far.
Third, TDs who don’t use pairing software can still just generate a quickie report (pretty much every word processing program of the last 10-15 years has the ability to save files as HTML), store it online, and display a link to it in MSA through the proposed field discussed two paragraphs above.
This is faulty reasoning Bill, and there are common examples in mathematics. Which gives you a better idea of the value of an event - to look at the probability of each portion an event and calculate an expected value - or to take either “0” or ALL the event as occurring?
Or for example - which would you rather do - if you are intending to hedge against a negative possibility, would you rather have an investment that is basically always the same - or an investment that varied along with the hedge?
The same concept is true here. Most events use standard tie-breaks. The 99% chance is that including standard tie-breaks will answer questions from participants. There is a 1% chance that it will create additional questions - AND EVEN THEN - the information is disclaimed and directs the player to the person who can answer the question.
If we can provide the correct answer MOST of the time, and minimize the risk for the times we can’t OF COURSE we should take this step.
Essentially, this is the 80/20 rule, only in this case, its probably more than 80/20.
**
Of course simply having organizers continue to do this is only a partial solution that doesn’t work well long term.
It is USCF - not organizers - that offer continuity and permanence. Records should be maintained and should be complete at the USCF level, not the organizer level. Organizers currently provide this service ONLY because USCF does not. Is there an organizer that would (routinely) post results, prize distributions, tie-breaks, etc. if USCF did? Of course not, we’d link to the USCF listing.
Part of USCF’s capability and part of the role that it offers is the role of record keeper. It only makes sense to centralize this role with USCF. Additionally, USCF should want to drive people to its website to strengthen ties and increase advertising opportunities which - if split up among organizers - are so small as to be basically nonexistent.
CLEARLY - this type of task belongs ONLY with USCF (long-term.) What I can’t understand is how people who have argued that it makes sense for USCF to take on a role that supports organizers are - in this thread - arguing that USCF should not do that.
We have to continue to move forward as an organization. This is one way to do that.
How is “a role that supports organizers” consistent with USCF calculating tiebreaks, which may differ from the tiebreaks used by the organizer?
I see major problems in using any tiebreak system not specified by the organizer.
Keep in mind, one way for the organizer to supply tiebreak points would be simply to upload the tiebreaks supplied by the pairing software. This would require virtually zero extra work for the organizer.
There could even be a checkbox, to be used in lieu of the above if the organizer so chose, to “please compute tiebreaks and list my crosstable in tiebreak order using the Modified Median system”. There could even be three such checkboxes, one each for first, second, and third tiebreak. Then, and only then, would it be necessary for the USCF software to actually compute the tiebreaks.
It’s consistent because organizers would not have to continue to post results that use standard tie-breaks.
It’s consistent because it would be clear that this may not be the same tie-break order.
Some organizers post .pto files, Bill. Without knowing the WinTD preference settings that doesn’t necessarily provide the same result either.
Your argument is basically that - since this doesn’t help 100% of the time, we should not do it, even though it may help 80%-99% of the time.
Sorry, I reject that argument. I think its silly. We need to take steps toward being better at things, and sometimes those steps will be imperfect.
It you want to volunteer to help, but it doesn’t help 100% of the time - do you simply reject doing it?
Good, but off-point, since no one is advocating that. What is being advocated is to provide the ability to sort a cross-table in the standard rulebook order and to post with that ability that this order may or may not be the order used by the organizer.
I agree!! I have said this all along. I’ve NEVER been against this. All I’ve said is that AS A FIRST STEP - UNTIL THE DATABASE CAN BE CHANGED AND MSA UPDATED, we can start by OFFERING an additional sort order, properly disclaimed. This takes relatively little work as it requires no additional data.
I completely agree and have advocated this all along. In fact, I’ve advocated updating the MSA fields to collect a wide range of data on a tournament that we don’t collect now so that MSA can correctly report results, and have advocated that as being the BEST WAY to support players and organizers.
It seems to me that you haven’t understood my proposal all along.
Being wrong 5% of the time is bad. This thread started with a complaint due to people not reading the existing note that the cross-table was not in tie-break order, and getting upset. That level of being upset would almost certainly go way up if a listed tie-break order did not match what the tournament used (thinking people would actually read a note, when they aren’t reading one now, seems naive).
A cost/benefit analysis has to balance the additional problems generated by being wrong against the benefits gained when right. So if we start with 20 people being mildly annoyed and that changes to 19 people not being annoyed and 1 person being furious, were things better off before or after the change?
Some would say do what you can to avoid mildy annoying the 19 while others would say that 1 furious person will blacken your reputation a heck of a lot more than 20 mildly annoyed people will.
It is only if it’s done the same way; it’s easy to solve. When you select the radio button for the T-B sort, just have a pop-up with the disclaimer and your agreement before showing the data.
And on what basis will one person be furious when he/she agreed that the order may not be correct?
I will take continuing to advance our website and continuing to improve communication to our members in better ways - including interactively getting their agreement - rather than sitting around for 10 years at a time and once again falling behind the times as we have done in the past.
Lead them through it and give them SOME credit, Jeff. Your argument above is basically - no matter how hard we try, people will be too stupid to understand it, so we shouldn’t try at all.
There might be a little too much overthinking of what the real problem is, and how to solve it. It occurs to me that the reason this is a problem is that USCF personnel and/or TDs and organizers have to field questions about why the list on the web site is not in the order of finish.
Perhaps the problem is with the phrase “not in tiebreak order”. The meaning of that phrase might not be as obvious as you think. Of course, after a moment’s thought, anyone who tried to figure out what it means would figure it out easily, but that misses the point.
When skimming through some introductory paragraphs like that, if people encounter a phrase they don’t understand, they will just breeze through it as if it didn’t exist. They won’t stop and try and understand it. They’ll just move on because, really, they aren’t all that interested in that part, anyway. They don’t care about that introduction, they just want to see what place their kid finished in. Never mind that the phrase in question directly addresses that very question. Because it isn’t a very natural phrasing of what the author meant, and it uses some insider terms, a typical reader won’t even realize he has read it.
It would be interesting to see if the rate of calls go down if you changed the wording on the web site to read, “Not in the order of finish.” or “Not shown in the order of final standings.” For emphasis, you might add that “the person shown in first place may not be the tournament winner.”