What is your preferred format for co-events?

[Ben Wilson]

It's been a while since we had this poll, and the recent bumping of the old 'preferred' format thread reminded me that there was a poll about this a while back, and with more co-events than ever, it'd be good to 'take the pulse of the community'.

Let me preface this by saying that there's plenty of room for all co-event types in the calendar, they all have their fans and there are positives and negatives to all options, which have been already debated to death on this board and no doubt will be again over and over.

The options:

Bristol format, also known as classroom format, is where everyone plays the same rounds as everyone else. One person hosts the game and there are two players to each table, letters are usually picked by one individual player each round for the whole room, while numbers can either work like that, or where each selection always consists of 4 large numbers and 6 small, and one player each pair chooses however many large numbers they want, with the small numbers coming from those to the immediate right of the large numbers. Everyone plays the same conundrum.

Edinburgh format, also known as Lincoln format, is where each table has three players on it- one host and two players. Everyone sat at the table gets a turn hosting, so a 'round' consists of three games. Each player picks letters and numbers only for their game, getting an equal number of round picks each game, and each game has an individual conundrum.

Finals stages- I've given three options down below. No finals stage is, as the name suggests, where there's no 'grand final' at the end of the day. Everyone plays the same set number of rounds, and whoever wins the most games/scores the most points is declared the winner at the end. A grand final is where only the top 2 play a grand final at the end of the set number of rounds, with the winner of that match being declared the winner. Division finals is where after the set number of rounds, players are grouped into 8 (or as near as possible where the number of attendees isn't an exact multiple of 8) and play a knockout tournament, with the winner of the top 8 being declared the winner of Division A, the winner of places 9-16 being declared the winner of Division B etc.

I've allowed everyone two options to vote as the bottom three options are technically separate from the top 6 (but hey, might as well kill two birds with one stone). The bottom three options are how fixtures are determined during the day.

Random/Random, as the name implies, means that all fixtures throughout the day are determined randomly, regardless of how a player has performed in previous rounds.

Random/Swiss means that the first round (first game/games for Bristol-style, first 2 games played per player for Edinburgh-style) is random, while the remaining rounds in the tournament will have people on a similar number of wins playing each other.

Seeded/Swiss means that the first round is drawn according to a seeding, whether this is Apterous Pro Ranks, on-screen performance or previous co-event performance. Generally, this means top players won't play each other in round 1.

Seeded/Random is also technically a possibility but I can't see any sensible reason why anyone would pick it.

This poll has no end date so feel free to discuss or change your vote at any time.

[JackHurst]

26 votes cast and 0 votes for either of the division based formats...

[George Armstrong]

I do get why some people like the divisions format as more people can win prizes not just those at the top of the tree, but IMO the big problem with the divisions format is that your final position is largely determined by how you perform solely in the first half of the day. Take the recent CO:NUT : Adam Latchford finished bottom Div A and won 3/7 games across that day, but because he was in Div A that guaranteed him 8th place. Fiona Titcombe came 1st in Div B and won 6/7 games across the day but because she was in Div B she could only finish 9th overall. I think if you made the overall final table (as in the one that's used for FOCAL standings overall wins then points, it would more fairly reward the better performances.

[Graeme Cole]

I do get why some people like the divisions format as more people can win prizes not just those at the top of the tree, but IMO the big problem with the divisions format is that your final position is largely determined by how you perform solely in the first half of the day. Take the recent CO:NUT : Adam Latchford finished bottom Div A and won 3/7 games across that day, but because he was in Div A that guaranteed him 8th place. Fiona Titcombe came 1st in Div B and won 6/7 games across the day but because she was in Div B she could only finish 9th overall. I think if you made the overall final table (as in the one that's used for FOCAL standings overall wins then points, it would more fairly reward the better performances.

Is a performance of 3 wins out of 7 necessarily that much worse than 6 wins out of 7 if for half the day the former player was artificially restricted to playing only the event's strongest opponents, while the latter player was prevented from playing those opponents?

The idea of the Swiss system is to use results from previous rounds to match players to suitable opponents. The divisioned-half-way-through format roughly approximates that idea but for some reason throws away half the data. If you've identified that divisions introduce their own unfairnesses, and you're proposing to rank the final standings as if it's all one big division again to eliminate unfairness, why stop there - why not get rid of the divisions altogether and then realise you've invented Swiss?

There's more to format preference than "fairness", of course. Arguably, Swiss 2-to-a-table Bristol format is objectively the fairest format in the list. Everyone gets the same selections, and the fixture generation for every game always takes into account *all* previous games, unlike the Lincoln format which has to generate your games two at a time. Nevertheless, more people prefer the Lincoln format, myself included.

So is there anything positive to say for the divisioned-half-way-through format? It's not the fairest, but do people enjoy playing it? Well, not according to this poll.

"But," asks a reader, helpfully deploying the question-from-the-audience narrative conceit, "C4C is full of high-rated Countdown nerds! Who reads forums nowadays anyway? Most Countdown viewers and eventgoers are on the Facebook groups now. This poll isn't representative of all skill levels. What about the better chance for lower-rated players to win a prize?"

It's an interesting point, but I don't think the current divisioned-half-way-through format is the answer. Firstly, most events already offer spot-prizes such as Tuff Luck, Overachievers, Nearest Points Total To A Random Number, Rudest Word, etc. Secondly, if you get the prize for finishing top of division C, what does that mean? You've finished top of 8 arbitrary players divided by performance based on three or four games one morning? Thirdly, the argument falls flat on its face when the Division B or Division C winner is someone who's good enough to be in Division A but got a bum draw in the opening games.

Nevertheless, if there's demand from lower-rated players for a divisioned tournament, what if it were divisioned *from the start* by rating or similar? I know of some chess and Scrabble tournaments that do/have done this. You have a number of divisions (let's say 3), and you assign players to divisions by rating, or some other long-term performance metric (Apterous rating? Previous event finishing positions, like for the FOCAL finals?).

For example, the organiser of a chess tournament might say "Division C is only for players rated under 1200, Division B is only for players rated under 1600, and Division A is for anyone." It's common to give players the option to play in a higher division than their rating, but not a lower one. Each division is its own separate tournament, with its own winner at the end of the day.

Advantages:

• Divides players by strength accurately, based on loads of previous games over weeks and months, rather than by three games that morning.
• You know the "Winner of Division C" prize is going to someone in the rating band it's designed for, not a higher-rated player who got stiffed by the draw and ended up in a lower division.

Disadvantages:

• There's no overarching ratings system for co-events.
• Not everyone has an Apterous rating or Pro Rank.
• Divisioned tournaments can only work well at all if there are a decent number of players. For an average co-event attendance you're probably looking at only two divisions (have I just invented the FOCAL Finals?)

[Gavin Chipper]

I think it's worth linking to this discussion because a lot of stuff has already been discussed so it's worth not having to reinvent every opinion twice.

Overall I think Edinburgh/Lincoln seems to work slightly better than Bristol for various reasons. It seems more sociable with three people looking after themselves on a table; stuff doesn't accidentally get given away by overhearing or clues to how easy a conundrum should be; you pick your own letters. Probably other stuff. But some people might find hosting games a little bit intimidating or boring or both. So there's pluses and minuses. I think variety can be good but overall I think a majority of Edinburgh/Lincoln tournaments is better.

I also think divisions aren't the best solution, which most people agree with, and a single grand final works well.

As for random/seeded/Swiss etc., I don't like seeded first rounds. I think random is better. Plus if the Swiss philosophy is to have similar people together then Lincoln's way of having a top, middle and bottom third player on each table in the first round and then switching to Swiss seems very clunky and illogical. You might as well just start with the top three seeds on a table and work down like that.

As for Swiss - it's a way of making sure that players on multiple wins have to play each other so you don't have loads of unbeaten players at the end of the day. It's quite unsatisfactory to have unbeaten players not making the final. However, it's not there for its fairness. A fair system would have people play the same average level of opponent over the day, whereas Swiss makes the top players play harder opponents and gives the lower down players weaker opponents, so it forces people towards the middle.

You could set up a system that makes unbeaten players play each other but other than that does it's best to keep players' average opponent level towards the middle. That would probably be best, but for simplicity you could have unbeaten players going against each other but the rest is just random. I'd favour either system over Swiss. But there's no option for these in the poll, so I'll abstain from that part. And I'll quote myself from the other thread (with a quote from Mark Deeks within that):

When people criticise this and that, one thing that often seems to escape criticism (except from me) is the Swiss pairs system itself. So I'll summarise:

1. It's intrinsically biased, favouring those who have done poorly so far. This is a mathematical fact.

2. It patronisingly assumes that people only want to play people of a similar level to them and can't possibly play one of the top players unless they are one themselves (except that in events like COLIN the first round is run to different logic, questioning the commitment to that philosophy anyway).

3. It means that weaker players are more likely to have the ignominy of playing Prune. It's annoying having to play Prune even if it's a free win. People don't want to play Prune and it's unfair to just foist it on the players who've done badly (which often ends up being the same people at many events) rather than spread it around.

And from Mark Deeks in this thread:

ADDITIONAL - At said Co:Lins, my favourite round is always the first one, where I'll play players I wouldn't normally.

[Fiona T]

Well I don't feel hard done by from the Newcastle tourney at all and almost certainly placed higher than I would have in a non-divisioned tournement, but I do think 4 games is far too few to then divide players based on performance. Division C was won by Tim Hebbes and division D by Bradley, both of who you'd expect to finish near the top of the field over the course of a non-divisioned tournament, and to whom I'm guessing focal points are a lot more important than they are to some of us who benefited from the early divisioning of the tournament.

I've attended a few scrabble tournaments and I'm really not a fan of enforced divisions which don't allow you to choose to play up, which also make it very hard to improve your rating to get out of those lower divisions - if you're towards the top you more or less have to win every game just to maintain your rating. The only reason at all I care about my scrabble rating is because it restricts who I'm then allowed to play with. I've had some great games (and taken some good scalps!) in tournaments with open divisions and much prefer that format.

[Philip A]

I've attended a few scrabble tournaments and I'm really not a fan of enforced divisions which don't allow you to choose to play up, which also make it very hard to improve your rating to get out of those lower divisions - if you're towards the top you more or less have to win every game just to maintain your rating. The only reason at all I care about my scrabble rating is because it restricts who I'm then allowed to play with. I've had some great games (and taken some good scalps!) in tournaments with open divisions and much prefer that format.

That’s why I played in the open division at the 2022 Nationals. No regrets. Gained some rating points.