In Turkey there is a dispute about the pairing (apparently made partially by hand) in the following tournament:

http://www.europeangodatabase.eu/EGD/To ... n=16350521

Maybe a bit international scrutiny from people not involved in the underlying debates might help.

* What do you notice?
* What would you do differently?
* What pairings are doubtful?
* If there are doubtful pairings, is there a pattern (to the benefit of which players?) in your opinion?

Ok, here's what I notice, and my opinion:

The bar is at 9 kyu, putting 13 players in the top group. Personally, I would've put it at 3 kyu, for 7 players in the top group, but this choice is not in itself unreasonable.

In the second round, there are 6 winners from the top group, which is the result of pairing down the weakest player (9k), who the lost (against a 12k). Better to pair down from the center of the field (2 or 3 kyu), in my opinion. It is more desirable to have 7 winners, because then you can avoid pairing the dan players amongst themselves in round two.

In round 2, a manual pairing was then made to still avoid pairing the four dan players amongst themselves, so two of them were paired down, while the other two were paired against the kyu players still within the top group. I like this, given the significant strength gap between the dan players and the rest of the field.

As a result, rounds three and four are a sort of KO between the four dan players, and effectively decide the tournament. This seems correct and desirable. The losing dan players from round three are paired in round 5, which is also effectively a place 3/4 play-off. Now personally, again given the large strength gap, I would've tried to arrange it so that the four dan players play round-robin in the last three rounds, but that would require even more manual pairings, and would really only have worked with a smaller top group and some heavy intervention in round 2.

All in all, I see no real problem with the top of the field.

Looking at the field in general, I notice that Hazar Tunca (7k) is higher up than expected at place 5, and Ender Tuccar (3k) is lower than expected at place 11. This, however, looks entirely reasonable given that the former has defeated the latter in round 3.

Other than that, the pairing for Onur Altinok might be considered slightly unfair. Especially being paired up again in round 3, after already being paired up due to manual intervention in round 2, is a bit off. On the other hand, if he were paired against weaker players in rounds 3 and 4, he would almost certainly have met a stronger opponent again in round 5. This pairing, had he won in round 3, would've actually given him more interesting opponents.

Overall, I see no significant problems?

I'm also a bit puzzled about why there's a problem. Is it that the top 4 only played each twice? You could argue that in five rounds they should play each other more often, but you're probably going to get cycles and tiebreaking that way. A KO is a good option I think. Also, its not just about the top 4 - the other players also deserve the most interesting possible tournament.

I also would have put the bar at 3k - I think the qualitative distance between 2d and 7k is very large, and the bar would have put the 7k players 4 MMS points behind the leading group. I do think there was an issue that Metin Ozsarfati and Onur Altinok ended up behind Hazar Tunca due to no real fault of their own, and the adjusted bar would have pretty much eliminated this problem I think.

In fairness, Hazar Tunca _did_ beat a 3k and only lose to dan players, and if his entry rank was 3k the results table would have looked entirely reasonable - my only intuitive objection is it looks like he got lucky, which isn't justified by the actual data.

Once the bar is set, I don't see anything particularly unreasonable about the draw itself, other than the variability you will get on what turned out to be more or less a 5 round swiss tournament (I exaggerate slightly there, but 70% of the tournament attendees being above the bar is _really_ high).

I'm quite surprised to see that handicaps were not used, even for players not in the top group. Is that usual?

As the original poster I am aware of the whole turkish discussion about the subject, still I would like to give me 2 cents.

I too wonder about the low bar, and the no-handicap (unavoidable after deciding the bar anyway) play of the tournament. I would probably put the bar at 3k and play with handicap reduced by 1 or 2.

On the top I actually like playing the tournament victory out in a kind of ko system among the four top players. Having them play 3 games against each other would be a very likely result in automatic pairing, but you would have a rather high likelihood of 2:2:2:0 or 2:2:1:1 ties - given the field these ties would then be resolved exclusively by SOS gained in pairings in lopsided games, i.e. tie breaking would be completely arbitrary. Given that pairing (and thus SOS points) was made manually in at least round 2 the result would then be wide open to objection. Doing the forced ko-tournament avoided this issue altogether, but on the other hand brought more lopsided games for the top players.

Another impression is that the pairing seems to consciously distribute games against the dan players more or less evenly among eligible kyu players (this might be the pairing software too). Some surprising pairings seem to me result of this decision, e.g. Kadir Sahin w/ only two wins paired against Bertan Bilen (4 wins at the point) in the final round, but Kadir Sahin had no games against the dan players at that point, while all better placed players with 2 wins had already. But given the settings, I believe this is a good thing to do.

I can't see anyone in particular benefitting from the alleged problems with the pairing. Although reducing the lopsidedness of games by a higher bar and some kind of handicap would have been better in my opinion - while the idea of playing a pseudo-k.o. tournament among the top players in this kind of field is imo a good one. Finally, player No. 18 has her name given wrongly in the EGD it should be Cagla Arslan.