Slightly prettier version of annotated game above:

http://gorgonian.******.com/chess-games.html

Aaaand we have a champion..... *drumroll* Congrats to Sholar!

How did that 'fish' (who blundered a rook in a silly way in round 1) manage to win the tourney?

Seriously though, congrats to the champion and to the runner-ups wlrs and DaMaGor!

In fact Sholar became the champion 2 months ago by drawing DaMaGor - neither wlrs nor DaMaGor could exceed Sholar's tiebreak.

Indeed, the final standings of round 4 in the top 3 are:

Sholar - 4 pts | tiebreak 11.5
wlrs - 3.5 | 7.5
DaMaGor - 3.5 | 7.5
leavenfish - 1 | 4

Sholar : wlrs = 1.5:0.5
Sholar : DaMaGor = 1.5:0.5 too
wlrs : DaMaGor = 1:1 (2 draws)

(Tiebreaks at chess.com are Neustadtl by default, see their explanation here.)

So Sholar's tiebreak would be 11.5 regardless of the result of the last finished game (DaMaGor vs wlrs, which was in fact the game for the silver). Both wlrs and DaMaGor had 3 points and tiebreaks 5.5 just before that game ended, and had one of them won it, he'd have 4 points with tiebreak 5.5 + 1*3 = 8.5 (= 0.5*4 + 1.5*3 + 2*1) < 11.5.

I thought I was doing well in my Black game vs. Sholar (though it might have just been a pretty looking equal position) until I forgot to include 31.-Bxa1 32.Rxa1 and instead played 31.-dxe2 straight away. boo @ me for regularly doing that in corr games.

Thanks!

Haha, yeah, that was painful. Also in round 2, with this game where I'm losing in the final (pre-repetition) position. I have no idea what I was doing there, just letting my queen wander and get trapped.

Yeah, I think by the end it was an equal position and the main danger had passed for me, but still complicated.

In the final round, especially, I was trying to play for more complicated positions (was inspired by a discussion, probably with Tex, YKW and wlrs, that basically amounted to: it's hard to beat stronger players with risk-free chess), and really got outplayed in that middle game there, and also by DaMaGor (again, following a good opening, I thought--a draw there ended the tournament).

I was pretty happy with my other game against you, first time in that opening for me, although I let another big opening edge slip there, too, and needed some luck to hold the draw.

Really enjoyed all of the games, so thanks to Noir_Desir for organizing. It's a bit of a joke that I've won this twice now, but as long as wlrs keeps forgetting that bishops can take rooks I still have a chance :-).

Congrats Sholar This is a great accomplishment given the playing strength of this forum.

So when does the next one start?

In any case, please ensure that it's two-round next time. There seem to be fewer active players than 2 years ago, so even 25-30 participants would make a great turnout.

If there's no predetermined number of participants, then 9(2)->2 seems to be a good catch-all option as long as there are 19-36 players (so 3 or 4 first-round groups).

For predetermined field sizes, the following options look good (TMI alert):

The format could be 4(2)->1 capped at 16-20, 5(2)->1 capped at 25-30 (ideally 25 so that groups consist of 5 people each, or 5 groups of 6 people each, but there should be room for 6 players in the final group in case of a first-round tie) or 10(2)->2 capped at 21 or 32 (i.e. with three 7-man and three-four 8-man groups and ideally a 6-8-man final group but room for 10 finalists in case of ties) or 8(2)->2 with 18 players (3 groups of 6) or 11(2)->2 with 24 players (3 groups of 8).

The time per move may be reduced from 3 days to 48 hours (a small trick ensuring that championship games are more likely to emerge at the top of personal ready game lists).

I don't know that I could handle that.

Yeah agree that 3 days is better, like the idea of bigger groups tho. I guess we can brainstorm away for a while in this thread and then set up the new edition.

3 days is definitely better -- two days is just not that much given that one can't accumulate any by playing faster. Bigger groups makes sense to me. And I can certainly see why wlrs would want to reduce variance in the final round :-)

In terms of the exact sizes, keep in mind the possibility for tied results resulting in 3 (or more) people advancing even from an N(2) system which can lead to an extra round.

Ties (with equal tiebreaks) are very rare. Look up any big tourney with groups of 6+, like the 27th or earlier 'Chess.com Tournament', and you'll see that 3-way ties happen less than 5% of the time. So it's fine to count on only 1-2 (but not always 0) extra advancing people per round.

I have nothing against 3 days per move. A cutoff based on the personal time per move stat usually does a better job of speeding up, but as the main candidates have ridic times per move like 15 or, in some cases, even 30 hours, I know this proposition is not going to pass

All right, I'm now ready to withdraw from a current private tourney (as I've just finished the last personal game of round 1) to make room for the 2+2 championship (non-premium members can be playing only in 1 private tourney at a time), so we can give it a go! (Lol, as if anything depends on me, just I've explained why I'm bumping.)

What's the expected number of participants? There are 20 players involved in at least one of the two team matches; perhaps the activity has dwindled since 2012, and there are a few non-playing celebrities to be invited and a few extra people who have taken part in the May-June TMs. On the other hand, some low-rated non-premium members whose chances for a medal are thin might choose to not take part here to enable playing in some other private (maybe thematic) tourney.

I think the 9(2)->2 format with a 27-player cap and a certain start date is fine. There will likely be 3 first-round groups of 6-8 players (I don't think more than 24 people will register), and the final group will likely consist of 6-7 players. If you don't like it that only 2 players per group will advance, make it 10(2)->3 with a 30-player cap (so that groups are 9-10 players strong, but there'll be a tad too many games at a time).

It's a good idea to first probe people's interest by making them register in a fictitional TM vs The Dummy Group (which is also useful for sorting group members by ratings).

CaptainPike's blog: how to create tournaments at chess.com, step-by-step What TDs can and can't do

Well, I think we got an "extra" round in this last iteration, but it was probably the result of just one tie.

Whoa. For me, 16+ simultaneous games would be a lot :-)

That the maximum number of entrants must be a multiple of the maximum group size is totally ******ed.

Maybe it can be circumvented by making an invite-only tourney with a fixed start date, groups of up to 8 players and a fictitional cap of 24, but sending merely 18 invites so that groups are normally 6-man (the final might be 7-8-man if there are ties).

Another option is 5(1)->1 capped at 20 players, maybe with some rating threshold like 1350+ (so that unserious players don't deny slow pony beasts their seats).

I think you mean 5(2) -> 1, coon74 -- I just looked this up, but if it's x(y) -> z, then it means a max of x players per group, y games played at a time, top z (and ties) players advance from each group. 5(1) -> 1 would mean only playing one game per opponent at a time, which would most likely stretch it out needlessly.

There was a tie in the 1st round, which made it 13 players to the second round, and another in the second round to make it 7 players to the third round. Had I not found a not-obvious only move in a R+P vs R ending against Sholar, it would have been 8 players -- two second-place ties out of three groups. So even if the tie that actually happened and extended the tournament hadn't happened, we came very close to having another that would have. (Then I tied wlrs in the final, but that didn't matter.)

If 24 is about the number we're expecting, 8(2) -> 2 capped at 24 seems good to me. Not excessively many games (14 isn't crazy, though it's getting kind of high), and pretty certain of having only two rounds. If we have fewer people, then that's fewer games in the first round.

Only problem with a rating threshold of 1350 (unless it's implemented by denying invites rather than by a rule in the tournament) is that it would seem to exclude new players (who start at 1200) -- Sholar was a new player for the first tournament and I was one for the second, for instance, and we might get others.

Yes, I meant 5(2)->1, sorry for the misprint.

Only two active TM players are now below 1200. So the invite filter will indeed work better.

You can absolutely start a tournament before it fills up, so no need to worry about the maximum number of total players.