The Pool Sucks

[Roger de Coverly]

This seems to only have happened in the past few years, so you could argue that the above may be a factor

The 4NCL is, I presume, using the "official" 8*ECF + 650. So there may be a case for reducing the 650 in order to get unrated players lower down the ranking list.

[Richard Bates]

I hope that views on the pools system are fed back to the 4NCL Board, incidentally. Although I don't know what would happen. It is much easier politically to move from 12 teams to 16 than to do the reverse (in the same way that it's easier to spend money recklessly on public services than it is to then cut back).

This means that extra care should have been taken before switching to the new system. Unfortunately the opposite was the case. There was no consultation about it and when the new scheme was announced a few years ago, it was coupled with an announcement about a new time control, and (it seems to me) that so many mails were swapped about that decision, which was eventually reversed, that virtually no one responded to this other announcement. In fact I might have been the only one who did respond to it.

Dare i say it, that some of us did point this out at the time, if not in an official capacity... (see (very) prior threads)

[Michael Yeo]

Brown Jack found themselves relegated last season despite having much better results against the same opposition as Anglian Avengers 1 and Wessex 1 who survived.

A bold statement considering Brown Jack currently sit 3rd with two teams just a point behind!

I'm guessing that you may have quoted the wrong bit of my post and that you were instead taking issue with :"Fortunately, it looks as though their (Brown Jack's) period in Division 3 will be short!"
Brown Jack have already played the 2 teams a point behind as indeed have Wessex 2. It is highly likely that the next 2 rounds will see the 2 Northern teams play Brown Jack and Wessex 2 which will be an interesting test of the relative strength of the Divisions. Probably Brown Jack v Wessex 2 will be Round 11. I think it will be very surprising if Brown Jack are not in the top 4 at the end.

[Roger de Coverly]

I guess this assumes that Atticus and Bradford will not be paired together in the next round.

As they've already played, twice if there were any pairings in the Jamboree, you would hope they were not paired. Personally I would have thought the rules should be that there would be no pairings between the Northern teams at Hinckley.

(edit) words added in case my original meaning was ambiguous

(edit) maximise pairings between Northern teams and the rest as stated in the rules seems to leave open a Northern v Northern pairing. I don't know why you would do this in the context of only 8 teams.

[Michael Yeo]

I guess this assumes that Atticus and Bradford will not be paired together in the next round.

As they've already played, twice if there were any pairings in the Jamboree, you would hope not. Personally I would have thought the rules should be that there would be no pairings between the Northern teams at Hinckley.

Rule 6.4 includes: "During this final weekend best endeavours will be made to maximise the number of matches played between Division 3 and Northern League teams" which I guess is the inverse of your thoughts.

[Joey Stewart]

You know, I just realised this arguement that it should only be matches against teams in 'your pool' falls flat when you consider that in division 3 the two top northern teams are allowed to carry points forward. It is good enough for them - so should it be for the rest of us who have had to work hard to earn our points.

[Mike Truran]

It's a completely different scenario. Divs 1 and 2 split into two. Div 3 and the NL combine.

[Mick Norris]

(edit) maximise pairings between Northern teams and the rest as stated in the rules seems to leave open a Northern v Northern pairing. I don't know why you would do this in the context of only 8 teams.

Because the rules were written when it wasn't known how many Northern teams there would be - if there had been 10, we wouldn't all have already played together - as it is, we have, so under swiss pairing rules, we shouldn't need to play again

[David Buckley]

16 teams, is too many for an 11 round APA and not really enough for an 11 round swiss so the present structure seems to be the least bad option, apart from changing the size of the divisions or number of rounds. Having said that, consider the following scenario: Team X is playing Team Y and Team X are 4-3 up. In the remaining game the Team X player has a probably drawn endgame in which he can indulge in lenghty winning attempts. He proceeds to do just that until it becomes clear that if he loses his game Team X and Team Y come third and fourth. If he wins or draws team X still come third in their pool and Team Y come fifth below Team Z. Since Team Z beat Team X in a previous round it is clearly in Team X's interests that their player lose his game. He returns to the board and 'fails' to notice that his opponent is threatening to fork his rook and his king.

Unfortunately there is nothing particularly contrived about this example. Sooner or later a situation is going to arise in one pool where a player is rewarded better for losing a game than winning or drawing. hopefully the player in question either won't be aware of it at the time or will care more about 'ethics' or his rating than how many points his team carries forward but it seems to me this is a serious flaw with the system as stands.

[Jonathan Arnott]

16 teams, is too many for an 11 round APA and not really enough for an 11 round swiss so the present structure seems to be the least bad option, apart from changing the size of the divisions or number of rounds.

I should have added that I think the strength of the 4NCL, Division One in particular, has declined or at least been severely diluted, as a direct result of the switch to a 16 team Division One.

I would have said it's the other way round, for male players at any rate. Previously you had 36 teams each looking for 7 players of 2000 standard and above. Now you only have 32 teams. The knock-on effect is that there are "spare" 2000+ players available to strengthen the 6 board division.

From the various quotes above and others, it seems there is a consensus that some problem exists:

1. Division 1's strength is diluted with 16 teams not 12.
2. The 'pool' system contains inherent unfairness, however the 'pools' are set up.
3. Division 1 has wide variation with too many results a virtual foregone conclusion (1 v 16 is likely to be a walk-over!)
4. Division 3 has teams which are as strong as most Division 2 sides.
5. 3 divisions of APA would be too much and leave the route from D4--->D1 to take too long.
6. 2 divisions of APA would push half of D2 into D3.
7. There aren't enough teams of quality to have a D1 and D2 Swiss as that would entail increasing to 40+ teams.

So let's look at a solution which 'solves' these problems, and ask whether the cure is worse than the disease...

(a) Make Division 1 a 12-team APA.
(b) This leaves 20 teams in D2 - not quite enough for a Swiss.
(c) Promote an extra 4 teams from D3 (or perhaps more logically don't relegate from D2) to make D2 a 24 team Swiss.
(d) Given the potential loss of teams from D3, combine D3 and the NL in Round 7 not Round 9.

My gut feeling is that this would be more 'stable' - ie there would be fewer teams playing in an inappropriate division. There might therefore be scope for slightly changing the way that promotion/relegation operates.

There would only be 4 'losers' from this scenario - those who miss out on playing D1 next season.

[Roger de Coverly]

There would only be 4 'losers' from this scenario - those who miss out on playing D1 next season.

Provided you give a season's notice ( implement from 2012 at the earliest), the pool system gives you a very simple implementation.

12 team division 1 is

8 from championship pool (edit)
plus
top 2 from div 1 relegation pool
plus
top 2 from div 2 promotion pool

[Mike Truran]

Doesn't that make 10?!

One of the problems with the old 12 team APA format was the way in which teams in or near the relegation zone regularly got hammered at the final weekend by teams packed with GMs for their matches against their championship rivals. The (bad) luck of the draw and the resulting collateral damage to teams unfortunate enough to be caught up in it frequently determined who went down and who stayed up.

[Roger de Coverly]

One of the problems with the old 12 team APA format was the way in which teams in or near the relegation zone regularly got hammered at the final weekend by teams packed with GMs for their matches against their championship rivals. The (bad) luck of the draw and the resulting collateral damage to teams unfortunate enough to be caught up in it frequently determined who went down and who stayed up.

You could, I suppose, do the pairings so that the final weekend featured seeds (1,2,3,4), (5,6,7,8) and (9,10,11,12) as mini groups.

Experience seems to show that the list of players as nominated back in September isn't always a good guide to the strength of the squad who actually play the matches. It would get tricky as well where one squad has two teams, particularly where the second team might finish in the top 4.

Perhaps you just do the seedings based on last year's results. That still leaves the problem of where to slot in the promoted teams. You could give them 10,11 and 12 but that's a problem if one of the new teams is likely to be a top four challenger.

[Jonathan Rogers]

Doesn't that make 10?!

One of the problems with the old 12 team APA format was the way in which teams in or near the relegation zone regularly got hammered at the final weekend by teams packed with GMs for their matches against their championship rivals. The (bad) luck of the draw and the resulting collateral damage to teams unfortunate enough to be caught up in it frequently determined who went down and who stayed up.

This happened to Barbican 2 as much as to anyone else between 2000-2003 (in the last weekend of 2000/1 we played Beeson Gregory 1 followed by Wood Green 1, each doing their level best to beat us 8-0). But we never complained about it. Why would we? We would have expected to lose to these teams earlier in the season anyway; and all the other teams in the division tended to be stronger in the last weekend too. It actually suited us to play most of our relegation rivals in earlier rounds.

To this day I still don't know exactly which relegation-threatened teams complained about uneven matches in the final weekend. Perhaps we should be thinking of teams playing Wood Green 2 or Guildford 2 in the last weekend, in which case they could well be playing quite different teams from those whom they would have played had they been paired at the start of the season. Even so, I cannot think of a very clear example when it determined anyone's fate.* That said, it is true that final weekend mismatches are best avoided and that more could have been done to try to avoid them, along the lines suggested by Roger.

Further thoughts:

The idea that a 16 team even would reduce mismatches always struck me as extraordinary, by the way. Under the old system, team number one would play teams two to twelve; and if they were able to beat team twelve by 8-0, well that's just too bad, and the problem rested with the lopsided way in which money is spent on players in the 4NCL. But why would it be better under the new pool system, when team one plays teams two to eight, and also something like teams nine, twelve, thirteen and sixteen?

It is to my mind self-evident that one can never seed teams satisfactorily at the start of the season - a conclusion I reached more than ten years ago. The problems have always been worst in the second division, incidentally. One can generally predict the top two or three teams in the first division, but more than once before the second division has been won by a complete outsider. I can still understand the desire to try to seed teams, but I have always been concerned about formats which put particular emphasis on predicting the strength/performances of the teams, and that was one of my many reasons for objecting to the new pool system.


* Late edit - probably Richmond in 2004/5 who played a colossal Guildford 2 side in the last round, needing to win. They played very well but could only draw, and went down. That was bad luck, then, to some significant extent attributable to the luck of the pairings. Still, this is only one example from a number of years (and note too that part of the problem here was the already identified one that it is hard to seed newly promoted teams, as was Guildford 2 in that season).

[Jonathan Arnott]

It is to my mind self-evident that one can never seed teams satisfactorily at the start of the season - a conclusion I reached more than ten years ago. The problems have always been worst in the second division, incidentally.

No real need to worry about seeding a D2 Swiss.

D1 can be seeded according to previous year's results, excepting the promoted teams.

Perhaps a rule should allow the arbiters discretion over where to slot the promoted teams in (discretion because it's disproportionately hard to quantify strength of a promoted team - based on things like average squad rating, position in D2 in previous season, average team rating in D2 previous season). To be honest getting the seeding 'right' is unimportant anyway. It's just making sure that you don't make any absolute howlers.