Your preferred format for CO-events

Discussion and announcements relating to unofficial Countdown competitions, held online or in real life. Observation, discussion, reflection, and other stuff ending in -ion.
Post Reply
Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Your preferred format for CO-events

Post by Gavin Chipper » Mon Dec 19, 2016 5:32 pm

There's been discussion over the years about what formats people prefer for CO-events, so I thought I might start a thread to see what people think. Obviously none of this is binding on CO-event organisers - I just think it might be interesting. There was a questionnaire that went round not that long ago, but that was more of a "secret ballot". I want an open and frank discussion!

There are a few variables, such as whether to have classroom "Bristol" style, where everyone plays the same rounds, or where you have groups of players (normally three) that have their own individual games - "Lincoln" style. Also some classroom events allow each table to have their own numbers pick (i.e. 1, 2, 3, 4 large or 6 small) rather than it being the same for everyone.

Then you have how the games are drawn. Some events give you random opponents, whereas in others you have "Swiss" pairs or triples, where you get to play other people who are in a similar position in the standings to you.

Some events just have "heat" games all the way through so your position is variable within the whole group throughout, and then with just a grand final at the end for the top two players. Others separate players into groups of 8 after a certain number of rounds. These groups have their own quarter finals, semi-finals and finals, so you are locked into a certain finishing position range of eight with three games still to play.

Then you also have the game format. Most seem to use 9-rounders, but sometimes you get 15-rounders, the old 14-round final format, or even 1/0 scoring where winning the points in the round gets you one point regardless of the length of your word.

There may be other stuff as well. Feel free to bring up what you want.

Anyway, over to you.

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Mon Dec 19, 2016 10:39 pm

I'm going to reply to my own thread.

On classroom-style v individual tables, I think both have their place. I probably overall prefer not having the hassle of hosting games, meaning a slight preference for classroom, but having both types of event works quite well I think.

With classroom-style, I think it's massively improved by individual games having their own numbers pick (started at Milton Keynes I think). And I think all classroom-style events should adopt this. I see it as a no-brainer. However, this works best when there are an even number of numbers games. At Huddersfield, there were three per game (in the 1/0 scoring rounds), so this made it unbalanced. I prefer formats that have an even number of numbers games generally.

While we're on game formats, I think the 9-rounders that most events use generally work quite well, although I have no problem with 15-rounders if there's enough time to still get a decent number of games in. 14-round finals I'm less keen on, partly because of the two conundrums. And it's not just because I'm shit at conundrums either. With the high level of competition at CO events, the top players struggle to take rounds off each other so crucial conundrums are very common. This means that their relative importance in deciding the games exceeds that of other rounds. One conundrum, fine - it's the big event at the end of a game. But two with one in the middle? Nah. And I kind of think that the 14-rounder was obsoleted by the new improved 15-rounder, which also has four numbers games.

The other format I've seen is the 10-rounder with 1/0 scoring. I think it goes LLNLLNLLNC. I find this a bit odd to be honest. First of all, I suppose the two things are separate. You can have 1/0 scoring with any round format, and you can have those 10 rounds with any scoring system. But as I said before, I don't really like formats with an odd number of numbers games. I don't have a massive problem with 1/0 scoring, but the events it's been used at have always had that for some games and normal scoring for others. And I find that a bit weird. I think it's best to stick to one.

Then we have the question of how to draw rounds - Swiss, random or something else. The Swiss pairs/triples puts people against players who have done similarly well up to that point - other than in the first round, where it can be random, seeded or whatever. But this is actually intrinsically unfair. It means that players who have done better get harder games than those who have done worse, rather than the fairest approach of everyone getting approximately the same level of opponents over the course of the day. This can lead to players who start badly leapfrogging over players who have done better in the last round because of the opponents that they have to face rather than because they've played better.

One advantage it does have is that it means that you're unlikely to have several unbeaten players at the end of the day. Two is OK because you can then just have a final. But if you have several, it's unsatisfactory to have people going home unbeaten but being defeated in the tournament. The "open" Birmingham tournament suffered from this because it had random draws and no knockout stage, and several people ended up unbeaten.

I think random draws are a bit fairer, although they can also be a bit of a lottery especially if there aren't many rounds. But while they can be a lottery, they don't intrinsically favour certain players like the Swiss system does. And because you end up with several unbeaten players, it then really needs a knockout stage afterwards with quarter finals etc., to make sure that you can't be unbeaten and still lose the tournament. But for time reasons that means fewer heat games. And at Bristol this year there were only three heat games. Not really enough to then lock people into a group of 8.

Some people like the Swiss system so that weaker players get to play other weaker players rather than spending the day getting thrashed. Personally, I think this reasoning is a bit patronising though. It's effectively a handicap system.

Maybe a better system still would be one that tries to match players so that everyone ends up with a similar average level of opponent over the day. But even then you could end up with several unbeaten players so it would probably require quarter final groupings. But with enough rounds, this shouldn't matter. Although Bristol only had three preliminary rounds, Huddersfield and St Albans ad more (though I can't remember how many).

Anyway, that will do for now. This post is long enough!

User avatar
James Laverty
Enthusiast
Posts: 364
Joined: Tue Jan 08, 2013 2:45 pm
Location: West Bridgford

Re: Your preferred format for CO-events

Post by James Laverty » Tue Dec 20, 2016 8:14 pm

I quite like how each event is different. Good variety of formats and allows each event to play to other peoples strengths
Definitely not Jamie McNeill or Schrodinger's Cat....

Matt Bayfield
Devotee
Posts: 521
Joined: Thu May 14, 2009 8:39 am
Location: Seated at a computer

Re: Your preferred format for CO-events

Post by Matt Bayfield » Wed Dec 21, 2016 9:20 pm

Nowadays, I prefer Co-events to be played on a board of 15 squares by 15 squares, with each player having their own selection of seven letters, which they must also manage in order to retain favourable selections for future rounds. Scoring should be based, in loose terms, on the obscurity of the letters in the English language, though I also like the idea of awarding bonuses for words declared in a pleasing geometric pattern. Numbers rounds should be confined to addition of numbers no higher than 10, and simple multiplication operations.

In the past, at the Co events I attended, I've always felt that the Bristol (classroom) style events comprising three randomly-drawn qualifying games, followed by a split into knockout groups of 8, produced some highly unpredictable placings in the midfield. I greatly admire the aim of inclusivity that such a format brings, allowing players to win prizes within their own "divisions", but all too often, the top prizes in lower divisions were claimed by players who received a horrendous draw in the random stage, leaving them on e.g. 1 win and in a division far lower than their "raw score" performance would merit.

In either classroom or head-to-head formats, Swiss pairings should consistently produce a fair (on merit) winner if the number of rounds is mathematically significant. Bristol-style events with 6 games of Swiss are fine for this purpose. (Five games, as previously used as CO:REA, is borderline.) However, in the Lincoln-style modified Swiss triplets over 6 games, pairings/tables are only re-drawn after every two games, and the insufficient mathematical significance of just 3 draws, is what causes the oddity Gevin describes - e.g. players on 1 win from 4 are able to make exaggerated gains in the final standings at the expense of players on, say, 3 wins from 4.

The algorithm for generating the modified Swiss triplets also works pretty poorly when the number of players is relatively small (18 or less). It leads to either repeat fixtures, or fixtures between opponents a large number of places apart in the standings.

Inevitably, whilst these are flaws of the currently used systems, I think we have to live with these. I certainly don't have any better ideas, as entirely random drawing for each round, over the small number of games played at Co events, should be statistically far worse in producing a "fair" result. (Obviously, if you want the results at your Co-event to reflect a greatly increased contribution of luck w.r.t. skill, then random draws every round are definitely the way to go.)

I don't like events which include showcase quarter-finals and semi-finals. People do generally enjoy a showpiece final, but most players attend Co-events to play as much as possible in the available time, rather than spectate.

The ability for players to opt between 6S/1L/2L/3L/4L in Bristol format is an excellent recent innovation, but I agree that it is best suited to games with an even number of the arithmetic rounds.

That'll do.

Heather Styles
Rookie
Posts: 90
Joined: Sat Oct 01, 2016 8:29 pm

Re: Your preferred format for CO-events

Post by Heather Styles » Thu Dec 22, 2016 12:24 pm

I'm personally getting quite fed up of 9-rounders. They're nowhere near as satisfying to play as 15-rounders. Hardly anyone from the 9-round era comes to events any more. I know 9-rounders mean more games can be played, but I'd rather play, say, four 15-round games in one day than six 9-round ones. I couldn't care less how numbers rounds are done, so long as it's as quickly and painlessly as possible. 14-rounders belong on the scrap heap of Countdown history. 1-0 scoring is probably the worst Co event idea ever. There is no perfect system of deciding who plays who, and as such perhaps there is something to be said for random pairings throughout. I don't pay go to events to spectate quarter- or semi-finals. I don't see the need for showpiece finals and don't particularly enjoy them.

Zarte Siempre
Devotee
Posts: 974
Joined: Sat Jun 30, 2012 8:56 pm
Location: Dadford, Buckinghamshire

Re: Your preferred format for CO-events

Post by Zarte Siempre » Thu Dec 22, 2016 2:59 pm

*shrug*

I'm mostly just happy to play Countdown and see people I like.

9s don't personally suit my game, but you fit more in with them, and they make it more unpredictable, which from an encouraging newbies point of view is fine. 1/0 scoring also achieves this a bit.

QFs and SFs are stupid (in a showpiece setting), but am fine with finals being showpiece. My only slightly issue with them is when it's 2 people who've already played. But I don't think you can go "We'll have a final, unless..."

But yeah, I just go with whatever as long as it isn't really obviously stupid. My priority is that I'm spending time with people I like, not getting too butthurt over how someone's decided to trial a new idea for an event. If it doesn't work, oh well - not the end of the world.
Possibly the first contestant to accelerate with a mic clipped...

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Thu Dec 22, 2016 3:23 pm

Zarte Siempre wrote:*shrug*

I'm mostly just happy to play Countdown and see people I like.

9s don't personally suit my game, but you fit more in with them, and they make it more unpredictable, which from an encouraging newbies point of view is fine. 1/0 scoring also achieves this a bit.

QFs and SFs are stupid (in a showpiece setting), but am fine with finals being showpiece. My only slightly issue with them is when it's 2 people who've already played. But I don't think you can go "We'll have a final, unless..."

But yeah, I just go with whatever as long as it isn't really obviously stupid. My priority is that I'm spending time with people I like, not getting too butthurt over how someone's decided to trial a new idea for an event. If it doesn't work, oh well - not the end of the world.
I think it's probably generally a good idea to go in with this attitude - that it's about enjoying the event and seeing people you like rather than worrying about the minutiae of how a competition is run.

But as it's come up - I think that QFs and SFs are OK if everyone is doing them. I agree that only the final should be a showpiece event. People don't come to spectate several rounds of other people playing.

Also, I've decided I don't really mind the Bristol lottery. For someone like me who tends to finish in the second half of the results (second half sounds better than bottom half), the more lotteristic (this is a word) it is, the more likely it is to be good for me!

User avatar
JimBentley
Legend
Posts: 2599
Joined: Fri Jan 11, 2008 6:39 pm
Location: Redcar, UK
Contact:

Re: Your preferred format for CO-events

Post by JimBentley » Thu Dec 22, 2016 6:41 pm

Zarte Siempre wrote:*shrug*

I'm mostly just happy to play Countdown and see people I like.
This is pretty much my view too. In fact I'd go further and say that it's the "seeing people" that comes way ahead of the "playing Countdown" bit for me. I've never been to any tournaments except Lincoln so can't really comment on the varying formats, but having said that, the classroom-style format kinda of puts me off as a lot of the fun at Lincoln is down to the format. Playing in groups of three - small enough so that everyone gets a good chance to talk - is a great way of getting to know new people.

As for the 9-rounder vs. 15-rounder thing, I would prefer 15 rounds as it places less importance on the conundrum, but for logistical reasons, I can't really see how this could be realistically be changed without extending the playing time to a seriously brain-hurting length, unless a tournament could be held over two consecutive days (and this would probably rule a lot of entrants out).

User avatar
Graeme Cole
Series 65 Champion
Posts: 1460
Joined: Tue Jul 06, 2010 9:59 pm

Re: Your preferred format for CO-events

Post by Graeme Cole » Thu Dec 22, 2016 8:24 pm

Lincoln/Bristol format: I'd play either, but I prefer Lincoln-style. I find it's more sociable and less like an exam or a classroom, and when there's a dedicated host for the game, conundrums work much better than the Bristol-style "look at the screen at the front, tap the table to buzz, and write the answer down, making sure to look at both the screen and the table at the same time so you can tell who buzzed first" thing.

Swiss/Random: Swiss. There's a reason they invented the Swiss pairs system, originally for chess tournaments, when doing random draws would have been easier. If you don't have enough rounds for a round robin tournament (which you don't, unless you have a really small number of players), Swiss will approximate the likely results of a round-robin tournament more accurately than random pairs. And for those who like the dramatic drawing-names-out-of-a-hat thing, with Swiss you can still do the first round randomly.

Matt makes a good point that even with Swiss, if you only have a small number of games this benefit is reduced. Also, with a Lincoln-style three-to-a-table tournament, it only makes sense to generate fixtures Swissly when everyone's played the same number of games, and that only happens after each player has played two games and hosted one, so when you play your second game of a round in a Lincoln-style tournament, that fixture has been generated without reference to the result of your first game. At COLIN, fixtures are only actually generated three times. These are certainly shortcomings of the Swiss three-to-a-table system, but I don't think a random draw would fix them.

This problem of "not enough games" also means it'd be a bad idea to switch to 15 rounders. COLIN and most other events don't use 9 rounders for historical or traditional reasons - Countdown had been using 15-rounders for years before COLIN started. 9 rounders are used so you can fit more games in.

The right answer would be to add more matches. You'd give a Lincoln-style tournament four rounds of three-to-a-table, so each player plays eight matches and hosts four. Timing constraints prohibit that though. COLIN would have to start at about 10am, which might have the negative side-effects of (a) some people who normally travel there and back on the day not coming, (b) less time for breakfast in the Little Chef, and (c) Jen missing the entire first round. So for Lincoln-style tournaments we're pretty much stuck with what we have unless someone invents something revolutionary that nobody's thought of yet. (5-round LNLNC games?)

If you've got a Bristol-style tournament, so two-to-a-table, with at least five or six games in it, I can't see a reason not to use Swiss. We've also got to remember that co-events should cater for players of all levels. At the Co-finals, when all the rounds were randomly drawn minus repeats, most of the newbies' first taste of co-events was losing seven out of seven when under Swiss pairs they'd have almost certainly played each other towards the end.

User avatar
Ben Wilson
Legend
Posts: 3671
Joined: Fri Jan 11, 2008 5:05 pm
Location: North Hykeham

Re: Your preferred format for CO-events

Post by Ben Wilson » Thu Dec 22, 2016 8:31 pm

JimBentley wrote:
Zarte Siempre wrote:*shrug*

I'm mostly just happy to play Countdown and see people I like.
This is pretty much my view too. In fact I'd go further and say that it's the "seeing people" that comes way ahead of the "playing Countdown" bit for me. I've never been to any tournaments except Lincoln so can't really comment on the varying formats, but having said that, the classroom-style format kinda of puts me off as a lot of the fun at Lincoln is down to the format. Playing in groups of three - small enough so that everyone gets a good chance to talk - is a great way of getting to know new people.
This, this, this, this, this. The appeal of co-events isn't playing Countdown- I can do that on apterous anytime I want. It's meeting up with the community and having a laugh. This is why I much prefer Edinburgh style over Bristol- you get to chat and piss about in Edinburgh, much less so in Bristol. It's also why I trialled the (phenomenally confusing) five to a table format at Nottingham a few years back- I figured that there's an optimum number of people at one table that generates optimum fun (I've had a long day at work, pardon my wording) that's probably higher than three. Or probably not.

The 9 rounder vs 15 rounder argument is simply one of logistics. I'd happily make COLIN all 15 rounders, but I doubt people would be willing to start playing at 10am and finish around 7pm.

Swiss pairs/triplets is the way I do my tournaments simply because of my background in Scrabble tournament organisation. It ensures that the cream rises to the top, and sure, there are some anomalies, but the only way around that is to level the playing field by playing a silly number of games. It also ensures people at the bottom of the table aren't going to play someone like Jack Worsley in the final round. And the occasional anomaly isn't a bad thing, surely? Would Gevin have finished 12th out of 36 at COLIN 2012 otherwise? :p So no, while I may experiment for the CoNot format a little (a lot) COLIN will largely be unchanged, though I'm still undecided over whether or not to do a completely random draw for round 1 rather than the usual seeded one. It seemed to work well enough at CoNot this year...

The only way COLIN has changed since 2005* also happened in 2012, funnily enough- the addition of the showcase final, and that was only thanks to a poll on c4c (can't find the link right now but it's in here somewhere) which asked for people's favourite format, and which 'Edinburgh style with a final' won fairly convincingly. Beforehand, the winner was the one who was top of the ladder after six games, and a final would only have happened if two players finished the day unbeaten (only happened once in twelve years). Don't see any reason to change that now.

tl;dr- come to COLIN. :)

*The only way apart from going to a seeded draw, as ironically enough, round 1 in 2005 had a random draw

User avatar
Thomas Carey
Devotee
Posts: 895
Joined: Sun Jan 09, 2011 4:17 pm
Location: North-West of Bradford
Contact:

Re: Your preferred format for CO-events

Post by Thomas Carey » Thu Dec 22, 2016 8:58 pm

Ben Wilson wrote:
The only way COLIN has changed since 2005* also happened in 2012, funnily enough- the addition of the showcase final, and that was only thanks to a poll on c4c (can't find the link right now but it's in here somewhere) which asked for people's favourite format, and which 'Edinburgh style with a final' won fairly convincingly. Beforehand, the winner was the one who was top of the ladder after six games, and a final would only have happened if two players finished the day unbeaten (only happened once in twelve years). Don't see any reason to change that now.

tl;dr- come to COLIN. :)

*The only way apart from going to a seeded draw, as ironically enough, round 1 in 2005 had a random draw
yeah
signature

User avatar
Matt Morrison
Post-apocalypse
Posts: 7167
Joined: Wed Oct 22, 2008 2:27 pm
Location: London
Contact:

Re: Your preferred format for CO-events

Post by Matt Morrison » Thu Dec 22, 2016 11:07 pm

Wow. That poll is brutal.

User avatar
Jennifer Steadman
Devotee
Posts: 955
Joined: Fri Jun 10, 2011 10:34 pm
Location: Kent
Contact:

Re: Your preferred format for CO-events

Post by Jennifer Steadman » Fri Dec 23, 2016 1:18 am

My preferred format for Co-events: basically exactly like Lincoln, but with a 15-round final. I like hosting games.

I'll play other formats/styles but really, really dislike changing formats constantly throughout the day, and not being able to pick my own numbers. Don't really like static groups either, but if there's been a lot of heat games then I can live with it.
proud tiara owner and annoying publicity person who tells you to click links to the the FOCAL website.

(and also to my TRAVEL BLOG)

Heather Styles
Rookie
Posts: 90
Joined: Sat Oct 01, 2016 8:29 pm

Re: Your preferred format for CO-events

Post by Heather Styles » Fri Dec 23, 2016 1:52 am

None of the details of game play bothers me very much, certainly not enough to stop me enjoying an otherwise enjoyable event. I respect all organisers' choices and innovations (even, begrudgingly, 1-0 scoring). To the 15-round doubters I would say it is possible to fit in five 15-round games without starting particularly early or finishing particularly late and while having a decent lunch break, as at Reading. But things are as they are. So here's to another successful year of Co eventing. Tl; dr come to COLIN and other events ☺

User avatar
James Laverty
Enthusiast
Posts: 364
Joined: Tue Jan 08, 2013 2:45 pm
Location: West Bridgford

Re: Your preferred format for CO-events

Post by James Laverty » Fri Dec 23, 2016 2:06 pm

While as I said earlier I like the variety, I'll give the more detailed account

Lincoln vs Classroom- Definitely Lincoln. I like getting the chance to host a few games as well as playing. I think the seeding works well at Lincoln; even if it can potenially leave newbies with a baptism of fire, the format means that after Round 1, you will generally be playing people at a similar level to you. My main problems with classroom are with the picking, although this has improved recently with the introduction of everyone picks their own numbers,it can be annoying when the letters pick goes the other way when fishing for a certain word. I also find the conundrums can be rather off-putting if everyone around has already got it and you're still looking, but tbf I can't think of a better system to use for it.

I think one or two of the criticisms of classroom are a bit harsh, particuarly the "randomly-drawn games can cause anomolies". It's all part of the fun and anomolies can easily happen in Lincoln style draws e.g. A player who wins one game in the opening round will be playing fellow one game winners who could range in ability. I vaguely remember being in a convo one year after COLIN (Can't remember who with or the year) talking about the possibilty of taking a "tactical defeat" in the first round to assure an easier draw for the rest of the day, and therefore possibly qualify for the final with 5 wins and a high total.

How Many Rounds- The main thing I'm in two minds about. I'm not a massive fan of playing 9rounders anymore, particuarly at Co-Events as they are very unforgiving and one mistake can be fatal, but they allow more games to be played and they are fair in that everybody picks 3 letters and a numbers round each. While I quite enjoy the "Perfect 10s" format as JR calls it, the extra numbers game at the end for Player 1 can often be vital for the result. I do like how games can end in a draw (as games can happen at MK), and if I was holding my own CoEvent I'd be tempted to allow this and only use tie-breaks for finals if needed.

I would personally rather play 15 rounders, but I don't think there would be enough time for them. Maybe have finals 15 rounds, with 9rs being played throughout the day. I'm not a fan of 14round finals and think the 15 would be a better compliment.

Finals/How many games- The main part of the classroom I like is the "finals" rounds in divisons of 8, as it feels everyone can build up to a finale of there own. One thing I would change though from CoHudd this year is the seeding through out, and keep it in a set bracket, instead of the current system where highest seed automatically plays the lowest remaining seed (Eg if 1, 2, 4, 6 were the QF winners, I'd keep it as 1 v 4 and 2 v 6 for the SFs, instead of 1 v 6 and 2 v 4 that happened this year).

As for finals at Lincoln style events, I think having a showpiece final is good, but JUST a final. Having SFs or even QFs for just the top 8 can be a bit disheartening for other players who don't get to play other games. I'll also say I find the Lincoln final a bit annoying when everyone crowds around a table watching, and it would be better if it could be done up at the front as a "showpiece" like at Classroom style events, but again I know this would be hard to to implement.

Other things- I've often found that the lunch breaks at CoEvents are a bit long, and could be cut down to about 40 mins, and if we started bang on 11.30 (or 12 if you're Jen), we might be able to squeeze a 4th round in. I'd personally prefer this as I think everyone playing 8 games would be more appropriate with it being a CoEvent and people could compare their 8 game runs as "octotals", although I think this might be unpopular with others and it may be seen as a bit tiring, but we managed 8 at MK as well.

Also, I quite like how many events have a little "side-challenge" like CoPointless or the Mole Hunt. I think these are good as 1) they are something a little different for the day, and 2) gives other people the chance to win a prize, so more of these :)

tl;dr, see you in Hykeham and all that.


*Just a thought, does Ben keep it at six as he was one of the few vicounts?
Definitely not Jamie McNeill or Schrodinger's Cat....

User avatar
Matt Morrison
Post-apocalypse
Posts: 7167
Joined: Wed Oct 22, 2008 2:27 pm
Location: London
Contact:

Re: Your preferred format for CO-events

Post by Matt Morrison » Fri Dec 23, 2016 2:45 pm

I'm not sure I've ever got back from a lunch break with more than -10 seconds to spare.

User avatar
Mark Deeks
Kiloposter
Posts: 1625
Joined: Tue Jul 12, 2011 3:15 am

Re: Your preferred format for CO-events

Post by Mark Deeks » Fri Dec 23, 2016 4:00 pm

To confirm much of the above, the threesome style (can we call it that? let's call it that) makes for a better experience socially, which is kind of the point. Also, some events go on for too long. It shouldn't be a full day. It should be a few hours with a lunch in between. I haven't played at a huge number of events because I keep turning up too late, but in my experience, the format is Co:LIN. Social, compact, and allows for both long nights, lie-ins and unnecessarily large breakfasts.
Eoin Monaghan wrote:
He may not be liked on here, but you have to give some credit to Mark

User avatar
Mark Deeks
Kiloposter
Posts: 1625
Joined: Tue Jul 12, 2011 3:15 am

Re: Your preferred format for CO-events

Post by Mark Deeks » Fri Dec 23, 2016 4:02 pm

ADDITIONAL - At said Co:Lins, my favourite round is always the first one, where I'll play players I wouldn't normally. It might be more fair and equitable, if there's a three tiered seeding system, to have only As play As and so on, but it's less fun, and fun really is the point. It's not like there's ever been an unjust winner via this system, apart from Corby that one time.
Eoin Monaghan wrote:
He may not be liked on here, but you have to give some credit to Mark

Matt Bayfield
Devotee
Posts: 521
Joined: Thu May 14, 2009 8:39 am
Location: Seated at a computer

Re: Your preferred format for CO-events

Post by Matt Bayfield » Fri Dec 23, 2016 5:13 pm

James Laverty wrote:I think one or two of the criticisms of classroom are a bit harsh, particularly the "randomly-drawn games can cause anomolies". It's all part of the fun...
We disagree here. I wouldn't describe the thrashings that result far more often from random draws, to be much fun for anyone. The most extreme example of these pitfalls of randomly-drawn qualifying games, happened at CoLei5 in 2012.

Dylan Taylor, whilst not yet at the peak of his ability, was already inside the top 20 of ProRanks, and the third highest ranked player in the field. In the three qualifying games, Dylan was randomly drawn against three of the other strongest players (the eventual winner, the eventual third place finisher, plus one other who I can't remember); lost all three games narrowly; and ended up in the bottom division. I can't imagine it was much fun for Dylan thrashing his opponents in that division for the rest of the day. Equally, I doubt his opponents enjoyed it much.

The inability to improve one's position after being placed in a division leads to quite some oddities. Theoretically, in a field of 32 players, player A could lose to three of the top players, then beat three of the weakest players, and finish 25th. A different player B could beat the same three weaker players, then lose to the same three top players, and finish 17 places higher in 8th.

Yes, three drawings of Swiss triplets has its failings, but spoiling players' fun isn't one of them - in my opinion.
Last edited by Matt Bayfield on Fri Dec 23, 2016 11:12 pm, edited 2 times in total.

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Fri Dec 23, 2016 5:40 pm

Graeme Cole wrote:Swiss/Random: Swiss. There's a reason they invented the Swiss pairs system, originally for chess tournaments, when doing random draws would have been easier. If you don't have enough rounds for a round robin tournament (which you don't, unless you have a really small number of players), Swiss will approximate the likely results of a round-robin tournament more accurately than random pairs.
This sounds like BS. Well, maybe it isn't but not because Swiss is in any way an accurate representation of a round robin, but because random might be even worse. Worse because it's more volatile and has more variance, but it has no intrinsic bias.

I've said this before, but a better system is probably to get people's "average" opponent to be the same level, or as close as possible. First round can be random. Then you rank all the players in position order and have a separate rank order in terms of average rank played. For example, if I've played ranks 4 and 8, my average is 6. Except that you're probably best taking out the player in question. So if I'm rank 20, then if I've played number 21, then on my list, they're number 20. Then in the each next round you put the player who's played the highest average rank against the bottom-ranked player, 2nd against 2nd last and so on. Do a bit of a fiddle so that players don't play each other twice and it's job done. It would be a bit more complex with triples though.

Swiss systems make the higher ranked players play higher ranked players and the lower ranked players play lower ranked players. It makes no effort to simulate a round robin. It's a handicap system that favours the worse players. I still would accept it for a tournament that has heats and then a single final, because you don't want to end up with several unbeaten players. But that's the reason it works. Not because it's fair, equitable, unbiased or anything. Because it's none of these.
Ben Wilson wrote:Swiss pairs/triplets is the way I do my tournaments simply because of my background in Scrabble tournament organisation. It ensures that the cream rises to the top, and sure, there are some anomalies, but the only way around that is to level the playing field by playing a silly number of games. It also ensures people at the bottom of the table aren't going to play someone like Jack Worsley in the final round. And the occasional anomaly isn't a bad thing, surely? Would Gevin have finished 12th out of 36 at COLIN 2012 otherwise? :p
Other systems throw out anomalies too!
JimBentley wrote:As for the 9-rounder vs. 15-rounder thing, I would prefer 15 rounds as it places less importance on the conundrum, but for logistical reasons, I can't really see how this could be realistically be changed without extending the playing time to a seriously brain-hurting length, unless a tournament could be held over two consecutive days (and this would probably rule a lot of entrants out).
This is basically my thoughts. Maybe an 8-rounder? (Leave out the conundrum! OK, maybe that wouldn't be popular.)
James Laverty wrote:One thing I would change though from CoHudd this year is the seeding through out, and keep it in a set bracket, instead of the current system where highest seed automatically plays the lowest remaining seed (Eg if 1, 2, 4, 6 were the QF winners, I'd keep it as 1 v 4 and 2 v 6 for the SFs, instead of 1 v 6 and 2 v 4 that happened this year).
I didn't notice that that had happened. I agree with you. I think it's really unfair that the number 8 seed would automatically have to play the highest seed left in the draw. If they defeat number 1, they should take their projected place in the draw.
Also, I quite like how many events have a little "side-challenge" like CoPointless or the Mole Hunt. I think these are good as 1) they are something a little different for the day, and 2) gives other people the chance to win a prize, so more of these :)
I've been to a couple of events where there has been a game of Pointless, and while I at first liked the idea in principle, I've since gone off it. I don't think it worked. It took up a lot of time, and most people were either not included or only included for a very short amount of time. At St. Albans, people got picked randomly to be in the game and everyone else was left out in the cold. And the game took a long time. Yes, it was quite fun to watch, but it took up too much time for something that wasn't very inclusive. At Huddersfield, I think the picking was even worse. There were team captains who picked their teams. So not only did some people not get to play, the reason was that they weren't popular enough or didn't know the right people, or people thought they would be shit. That's not nice, and is reminiscent of PE lessons at school. It was a very bad idea. And although I did get on a team, I still didn't get to play because my team got knocked out before I got a turn (Ryan Taylor went for some stupid guess when he could have just played safe - dickhead). So two (long) games and many people didn't get to play at all. So a nice idea in principle, but it doesn't work in practice for the good of everyone.
Mark Deeks wrote:Also, some events go on for too long. It shouldn't be a full day. It should be a few hours with a lunch in between. I haven't played at a huge number of events because I keep turning up too late, but in my experience, the format is Co:LIN. Social, compact, and allows for both long nights, lie-ins and unnecessarily large breakfasts.
I agree that some events start too early. 10:00 is too early. 11:00 is the earliest an event should start in my opinion.
Mark Deeks wrote:ADDITIONAL - At said Co:Lins, my favourite round is always the first one, where I'll play players I wouldn't normally.
I agree with this. It's another reason not to have the Swiss system. I like the idea that I could play anyone, rather than ending playing the same people at each event after round 1. Swiss is an intrinsically less sociable system! But, still, it does ensure you don't get several unbeaten players.

User avatar
Jennifer Steadman
Devotee
Posts: 955
Joined: Fri Jun 10, 2011 10:34 pm
Location: Kent
Contact:

Re: Your preferred format for CO-events

Post by Jennifer Steadman » Fri Dec 23, 2016 6:52 pm

James Laverty wrote:I vaguely remember being in a convo one year after COLIN (Can't remember who with or the year) talking about the possibilty of taking a "tactical defeat" in the first round to assure an easier draw for the rest of the day, and therefore possibly qualify for the final with 5 wins and a high total.
This always comes up as a supposed failing of the system but has anyone ever actually done this at a Co-event? Surely everyone's too proud, and most people are too honest. As a 'loophole' it's reliant on a bunch of arrogant and complacent assumptions anyway.
proud tiara owner and annoying publicity person who tells you to click links to the the FOCAL website.

(and also to my TRAVEL BLOG)

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Fri Dec 23, 2016 7:01 pm

Jennifer Steadman wrote:
James Laverty wrote:I vaguely remember being in a convo one year after COLIN (Can't remember who with or the year) talking about the possibilty of taking a "tactical defeat" in the first round to assure an easier draw for the rest of the day, and therefore possibly qualify for the final with 5 wins and a high total.
This always comes up as a supposed failing of the system but has anyone ever actually done this at a Co-event? Surely everyone's too proud, and most people are too honest. As a 'loophole' it's reliant on a bunch of arrogant and complacent assumptions anyway.
Even if no-one's actually done it on purpose, the fact is that someone can lose a game, and end up doing better as a result. I think people have said that Chris Wills won COLIN overall from the second table in the final round because he had easier games. This was in the days before there was a grand final.

User avatar
Innis Carson
Devotee
Posts: 876
Joined: Sat Nov 15, 2008 3:24 pm

Re: Your preferred format for CO-events

Post by Innis Carson » Fri Dec 23, 2016 7:53 pm

Gavin Chipper wrote:Even if no-one's actually done it on purpose, the fact is that someone can lose a game, and end up doing better as a result. I think people have said that Chris Wills won COLIN overall from the second table in the final round because he had easier games.
That seems rather speculative. Who can say categorically that the 'pros' of Chris losing an earlier game (theoretically easier opponents in the last round which he was theoretically more likely to beat, and score highly against) outweighed the 'cons' (definitely no longer being able to afford losing one of his remaining games, and forfeiting the chance to force victory regardless of score by winning both)? It probably is true in some cases, but I think it must be rare enough that this can concretely be stated to be the case that it represents only a fairly minor weakness in the format, and that it would be a very foolhardy strategy indeed to try something like this on purpose. Especially nowadays where there's the very real possibility of two players finishing the heats undefeated and you being left feeling like an idiot.

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Fri Dec 23, 2016 11:28 pm

Innis Carson wrote:
Gavin Chipper wrote:Even if no-one's actually done it on purpose, the fact is that someone can lose a game, and end up doing better as a result. I think people have said that Chris Wills won COLIN overall from the second table in the final round because he had easier games.
That seems rather speculative. Who can say categorically that the 'pros' of Chris losing an earlier game (theoretically easier opponents in the last round which he was theoretically more likely to beat, and score highly against) outweighed the 'cons' (definitely no longer being able to afford losing one of his remaining games, and forfeiting the chance to force victory regardless of score by winning both)? It probably is true in some cases, but I think it must be rare enough that this can concretely be stated to be the case that it represents only a fairly minor weakness in the format, and that it would be a very foolhardy strategy indeed to try something like this on purpose. Especially nowadays where there's the very real possibility of two players finishing the heats undefeated and you being left feeling like an idiot.
I don't think anyone probably would do it on purpose, but the Swiss system works to shove people towards a crowded centre, and people will swap places when they might otherwise have had a gap between them - not just for first place. I'm more commenting on what the Swiss system can do anyway, not whether people would actually try to game it in this way.

User avatar
Ben Wilson
Legend
Posts: 3671
Joined: Fri Jan 11, 2008 5:05 pm
Location: North Hykeham

Re: Your preferred format for CO-events

Post by Ben Wilson » Fri Dec 23, 2016 11:46 pm

Innis Carson wrote:
Gavin Chipper wrote:Even if no-one's actually done it on purpose, the fact is that someone can lose a game, and end up doing better as a result. I think people have said that Chris Wills won COLIN overall from the second table in the final round because he had easier games.
That seems rather speculative. Who can say categorically that the 'pros' of Chris losing an earlier game (theoretically easier opponents in the last round which he was theoretically more likely to beat, and score highly against) outweighed the 'cons' (definitely no longer being able to afford losing one of his remaining games, and forfeiting the chance to force victory regardless of score by winning both)? It probably is true in some cases, but I think it must be rare enough that this can concretely be stated to be the case that it represents only a fairly minor weakness in the format, and that it would be a very foolhardy strategy indeed to try something like this on purpose. Especially nowadays where there's the very real possibility of two players finishing the heats undefeated and you being left feeling like an idiot.
On the two occasions when Chris won COLIN with 5 wins (2006 and 2007) his losses were to Kirk Bevins and Paul Howe by 5 and 4 points respectively, so i think we can safely knock that rumour on the head.

Phil Collinge
Acolyte
Posts: 148
Joined: Sat Mar 28, 2009 11:33 pm
Location: Southend-on-Sea via Burnley

Re: Your preferred format for CO-events

Post by Phil Collinge » Mon Jan 23, 2017 1:56 am

Ben Wilson wrote:
Fri Dec 23, 2016 11:46 pm
Innis Carson wrote:
Gavin Chipper wrote:Even if no-one's actually done it on purpose, the fact is that someone can lose a game, and end up doing better as a result. I think people have said that Chris Wills won COLIN overall from the second table in the final round because he had easier games.
That seems rather speculative. Who can say categorically that the 'pros' of Chris losing an earlier game (theoretically easier opponents in the last round which he was theoretically more likely to beat, and score highly against) outweighed the 'cons' (definitely no longer being able to afford losing one of his remaining games, and forfeiting the chance to force victory regardless of score by winning both)? It probably is true in some cases, but I think it must be rare enough that this can concretely be stated to be the case that it represents only a fairly minor weakness in the format, and that it would be a very foolhardy strategy indeed to try something like this on purpose. Especially nowadays where there's the very real possibility of two players finishing the heats undefeated and you being left feeling like an idiot.
On the two occasions when Chris won COLIN with 5 wins (2006 and 2007) his losses were to Kirk Bevins and Paul Howe by 5 and 4 points respectively, so i think we can safely knock that rumour on the head.
Worth noting also that Chris Wills took part in 6 of the first 8 Co-Events, won his first five and finished 3rd in the other.

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Tue Jan 24, 2017 5:29 pm

Gavin Chipper wrote:
Fri Dec 23, 2016 5:40 pm
I've said this before, but a better system is probably to get people's "average" opponent to be the same level, or as close as possible. First round can be random. Then you rank all the players in position order and have a separate rank order in terms of average rank played. For example, if I've played ranks 4 and 8, my average is 6. Except that you're probably best taking out the player in question. So if I'm rank 20, then if I've played number 21, then on my list, they're number 20. Then in the each next round you put the player who's played the highest average rank against the bottom-ranked player, 2nd against 2nd last and so on. Do a bit of a fiddle so that players don't play each other twice and it's job done. It would be a bit more complex with triples though.
I was thinking about this a bit more and I've decided I have a better system for deciding on pairs. Swiss pairs is used in competitions because it stops there being too many unbeaten players at the top (by forcing them to play each other), so with the right number of players/rounds, it effectively works as a knockout tournament for the winner while still giving other people games to play. But obviously it has the problem that players who do better get harder draws so it's biased in that sense.

But I think a good compromise would be for the system to put the unbeaten players against each other but then use what I've described above for all other players. But among the unbeaten players, there's no need to put 1 & 2 against each other, and 3 & 4 etc. It could still use the above system but with the restriction that unbeaten players face other unbeaten players.

The exception would be when there are an odd number of unbeaten players. In that case, I'd probably just also add the next player down into the "top pool" as well. So if there are 7 unbeaten players, then the top 8 will all face someone else from the top 8. Then of course if someone goes a few rounds unbeaten and gets really hard opponents, once they lose a game, their next opponent is likely to be someone ranked quite low down to balance things out.

As above, triples would be complicated, but I think this would work very well for all "pairs" tournaments and should be adopted immediately. :ugeek:

Ryan Taylor
Postmaster General
Posts: 3661
Joined: Sun Oct 25, 2009 6:18 pm

Re: Your preferred format for CO-events

Post by Ryan Taylor » Tue Jan 24, 2017 7:05 pm

I prefer playing Bristol-style rather than Lincoln style because then I can be assured that the letter distribution is the same for everyone and that the letters are being shuffled properly.

Matty Artell
Rookie
Posts: 46
Joined: Mon Nov 10, 2014 9:38 pm

Re: Your preferred format for CO-events

Post by Matty Artell » Sat Feb 18, 2017 6:49 am

Matt Morrison wrote:
Thu Dec 22, 2016 11:07 pm
Wow. That poll is brutal.
Interestingly differs significantly from my findings - I did a survey about a year ago which received 43 responses - 53.5% in favour of Lincoln and 46.5% in favour of Bristol.

User avatar
Jennifer Steadman
Devotee
Posts: 955
Joined: Fri Jun 10, 2011 10:34 pm
Location: Kent
Contact:

Re: Your preferred format for CO-events

Post by Jennifer Steadman » Sat Feb 18, 2017 7:32 pm

Matty Artell wrote:
Sat Feb 18, 2017 6:49 am
Matt Morrison wrote:
Thu Dec 22, 2016 11:07 pm
Wow. That poll is brutal.
Interestingly differs significantly from my findings - I did a survey about a year ago which received 43 responses - 53.5% in favour of Lincoln and 46.5% in favour of Bristol.
53.5% is an overwhelming majority these days. All events should henceforth be Lincoln style.
proud tiara owner and annoying publicity person who tells you to click links to the the FOCAL website.

(and also to my TRAVEL BLOG)

Heather Styles
Rookie
Posts: 90
Joined: Sat Oct 01, 2016 8:29 pm

Re: Your preferred format for CO-events

Post by Heather Styles » Mon Feb 20, 2017 11:06 am

Ryan Taylor wrote:
Tue Jan 24, 2017 7:05 pm
I prefer playing Bristol-style rather than Lincoln style because then I can be assured that the letter distribution is the same for everyone and that the letters are being shuffled properly.
I agree. This is another advantage of Bristol style.

User avatar
Matt Morrison
Post-apocalypse
Posts: 7167
Joined: Wed Oct 22, 2008 2:27 pm
Location: London
Contact:

Re: Your preferred format for CO-events

Post by Matt Morrison » Mon Feb 20, 2017 12:24 pm

Overall I was pleasantly surprised by the Bristol format. I think Lincoln edges it slightly for me still, mostly because the fact that conundrums have to be made really hard in order to not ruin the elite games which made them pointless in every game I played. (I might be wrong in remembering this but I think in the seven games I played there was only one buzz.)

But yeah it's good that Jono's cheating is counteracted - then again as this was computer-based, I thought the algorithm would be Countdown-style, and not allow for the freak rounds we had a couple of times where we had 4 of the same consonant/vowel and so on.

Gavin Chipper
Post-apocalypse
Posts: 7675
Joined: Mon Jan 21, 2008 10:37 pm

Re: Your preferred format for CO-events

Post by Gavin Chipper » Mon Feb 20, 2017 1:14 pm

Matt Morrison wrote:
Mon Feb 20, 2017 12:24 pm
But yeah it's good that Jono's cheating is counteracted
A two-year CO-event ban would have been a safer way.

Matt Bayfield
Devotee
Posts: 521
Joined: Thu May 14, 2009 8:39 am
Location: Seated at a computer

Re: Your preferred format for CO-events

Post by Matt Bayfield » Mon Feb 20, 2017 6:01 pm

Matt Morrison wrote:
Mon Feb 20, 2017 12:24 pm
But yeah it's good that Jono's cheating is counteracted - then again as this was computer-based, I thought the algorithm would be Countdown-style, and not allow for the freak rounds we had a couple of times where we had 4 of the same consonant/vowel and so on.
My Excel spreadsheet (which is the one used at Bristol to pick and display the Letters) shuffles the C and V packs into completely random orders, so whilst the Letters distribution matches that of apterous (circa 2013), and is therefore kinder than most box games, it's still possible to get three Us in a row, for instance. The shuffling macro wasn't ever intended to be sophisticated enough to separate clumps of identical letters. My spreadsheet was merely a freely-offered solution to player requests to:
(a) ensure that a favourable Letters distribution is used, rather than box games with way too many Us, two Js, three Ws, etc.
(b) avoid rounds which are very similar to previous rounds displayed that day, because of poor human shuffling of the C and V packs.

If anyone wants to adapt my spreadsheet to do some sort of "face up" shuffling which separates clumps of the same Letters, or clumps of Letters which combine poorly, I have no problem with that. It shouldn't be particularly hard to do. (Although if anyone does do this, I'd like a copy, please!)

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest