Page 1 of 1

FOCAL scoring system

Posted: Sun Aug 12, 2018 11:02 pm
by Tim Down
I’d like to propose a new scoring system for FOCAL standings from 2019 onwards.

The current system works fine when there are similar attendances at each event. However, a particularly large event such as CO:LON, which has roughly double the number of players of an average event, carries so many points that it unjustifiably dwarfs a good result in a smaller event. For example, in 2018 when CO:LON was attended by 67 players and CO:Leeds by 33 players, first place at Leeds scored the same number of FOCAL points as 35th place at London. Given the quality of field at Leeds, it’s difficult to argue that these represent similar levels of achievement. I do think it’s reasonable to award a few more points for the larger events, however, so my proposal is as follows:

Each event has a maximum number of points (call it N) that can be scored that is determined and published at the start of the FOCAL year. The winner is awarded the maximum number of points (N) and everyone else’s points are distributed evenly between 0 and N based on their position. For example, if an event is worth 500 points and has 25 players, the winner gets 500 points, second gets 480, third 460 and so on down to 25th who gets 20.

The rest is a just a matter of determining how many points to award for each event. I suggest using numbers significantly higher than the number of players at the event to reduce the effect of rounding.

I’ve created a web page showing the current and proposed scoring systems, applied to the 2018 results so far. It lets you tweak the points available for each event. In my example, there are three tiers of events: tier 1, worth 600 points (COLIN, CO:LON, .co.mk), tier 2 worth 500 points (COLIN Hangover, CoBris, Co:Leeds, CoCam) and tier 3 worth 400 points (CoNuT, Co:Wat). I’ve based this on attendance in 2018. It means that larger events are still worth more than smaller ones but to a lesser degree than at present.

http://timdown.co.uk/code/countdown/focal.html

Notes

  • I couldn’t find a list of placings from all events so mine are calculated from http://greem.co.uk/cgi-bin/coevents/event.py. My numbers don’t exactly match the most recently published standings (of which I only have the top 8) so I’ve clearly got something wrong somewhere. They’re close though. (EDIT: Graeme has now kindly furnished me with accurate placings so I'll update the page soon.)
  • My own position is higher in my proposed scoring system. However, this isn’t a big surprise because what prompted me to look into this was my dismay at how much ground I’d lost after being unable to attend CO:LON. In any case, I’m not proposing changing the scoring for 2018 so I wouldn’t benefit anyway.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 12:32 am
by Graeme Cole
I agree that the current scoring system does reward the top positions at smaller events less than it should compared to the middling positions at large events.

However, I'm not convinced that the points value of an event (number of points given to the winner) needs to be decided in advance of the event. All that means is that someone has to pluck a number out of the air based on expected turnout many months in advance, and/or some subjective view of how "prestigious" the event is. Why guess? I think it's worth keeping the direct link between the number of players and the number of points available at an event.

However, there's no reason why it has to be a linear relationship, as it is now.

You could come up with particular curve on a graph, something of the form y=ax^k, where 0 < k < 1, and a is some constant multiplier to give the y-value the desired scale. (If anyone is having trouble visualising what that actually looks like, it's a curve that starts at zero, goes up sharply to start with, then levels off. It's always going up, but at a progressively shallower gradient.)

The parameters of the curve could be chosen such that, say, the points value of a 40-person event was significantly more than the points value of a 20-person event, but the value of a 60-person event was not-so-significantly more than that of a 40-person event. So winning a larger event would still score more than winning a smaller one, but FOCAL scores would no longer be so heavily skewed by results (or non-attendance) at larger events.

One other possible avenue of exploration is to de-linearise the relationship between finishing position and points awarded - give the first placed player (say) 500 points, the second-placed player 300, the third 200, then work linearly down from there, with the intention that first place at a small event would outscore most middling positions at larger events. (Edit: did Gevin suggest something like this a while ago?) However, I haven't run any numbers on this so I have no idea whether it'll produce results that make sense. For that reason I'm hesitant to suggest that as an "improvement" based on my own vague 1am guesses when I know that Zarte did a lot of work reviewing simulations of various scoring systems on previous years to come up with the system we've got now. One problem I can already see with it is it might make the top players even harder to catch up with.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 7:27 am
by Thomas Carey
viewtopic.php?f=10&t=11281&p=165157#p162047

This was Gev's thing. Quite liking the idea of your first suggestion atm Graeme (would need to run it with past events obvs)

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 8:09 am
by Zarte Siempre
As Graeme said, several systems were trialled pre-FOCAL.

It was important to us that good players who only turn up occasionally and thus were unlikely to come to Birmingham weren't what ended up in the top 8. That, for lack of better term, tier B players who support the events regularly had a shot at being rewarded for that. That the mechanic was simple to explain to anyone who happened to ask, to the point where they could reasonably work out exactly how they'd done on the day. And equally that it would be easy to follow in terms of looking over the spreadsheet. For this reason, things with a curve (except perhaps a consistent pre-defined one like the trial we did using... think it was some sort of motor racing scoring) were definitely out.

Personally I believe as well that organisers where possible should be incentivised to work harder to promote their event, knowing that some people will base their decisions on where they can take points from. Additionally, I believe that larger events tend to (although there are some exceptions) have much harder fields to get good results out of, and can easily swallow very good players up meaning I don't believe that the difference is as minimal as people seem to think. The issue with this is not only does it over-equalise the winning of some events IMO, but tiering the events also removes much of the incentivisation as the boundaries become so wide. There is a big difference between looking for 5 extra people to tip the points value up by 10, to having to find 10, and if you do, having no need to look any further

Additionally, although much more minorly in the grand scheme, a simple system allows me to extrapolate patterns much more easily, meaning my post-event summaries take me not too much time to do. I'm not committing to continuing them if that's going to make my head hurt on a co-event Sunday when my head often already hurts :)

I get the issues people have with the current system, and I get that it's not perfect. But I felt that it was the right one at the time, and I'm yet to see one that eradicates enough issues to make it worth changing it when it's guaranteed to lose much of the simplicity and self-calculation. Ultimately though it's not my call, it's Jeff's. Appreciate that you've put thought into this though, and apologies I'm not glowing about it.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 8:31 am
by Gavin Chipper
Reading with interest to see how this develops.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 9:11 am
by Tim Down
Zarte Siempre wrote: Mon Aug 13, 2018 8:09 am I get the issues people have with the current system, and I get that it's not perfect. But I felt that it was the right one at the time, and I'm yet to see one that eradicates enough issues to make it worth changing it when it's guaranteed to lose much of the simplicity and self-calculation. Ultimately though it's not my call, it's Jeff's. Appreciate that you've put thought into this though, and apologies I'm not glowing about it.
That's fine, no apology needed. I'm not at all precious about my proposal and I'd read the old threads and knew what to expect. Fundamentally, nothing's changed and I'm just bringing it up again now that it actually affects me. I didn't look beyond a linear relationship between event placings and FOCAL points because of the desire for simplicity expressed by you (and others?) but I'd be just as happy with something more complicated.

I think the problem I outlined (too many points for middling finishes at big events) is serious enough to warrant change and I'd prioritise fixing it over simplicity. However, I don't have to deal with people asking about the scoring system. Does it happen a lot? Finally, I enjoy your post-event summaries and I'd be sad if they stopped.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 9:26 am
by Zarte Siempre
Tim Down wrote: Mon Aug 13, 2018 9:11 amI think the problem I outlined (too many points for middling finishes at big events) is serious enough to warrant change and I'd prioritise fixing it over simplicity. However, I don't have to deal with people asking about the scoring system. Does it happen a lot? Finally, I enjoy your post-event summaries and I'd be sad if they stopped.
A reasonable number of people when they first turn up ask about it, and then again when they first start attending enough for it to be relevant. And then again when they sober up. And in some cases again when they couldn't be bothered to listen properly the first times.

Re: FOCAL scoring system

Posted: Mon Aug 13, 2018 3:54 pm
by Gavin Chipper
Zarte Siempre wrote: Mon Aug 13, 2018 8:09 am Personally I believe as well that organisers where possible should be incentivised to work harder to promote their event, knowing that some people will base their decisions on where they can take points from. Additionally, I believe that larger events tend to (although there are some exceptions) have much harder fields to get good results out of, and can easily swallow very good players up meaning I don't believe that the difference is as minimal as people seem to think. The issue with this is not only does it over-equalise the winning of some events IMO, but tiering the events also removes much of the incentivisation as the boundaries become so wide. There is a big difference between looking for 5 extra people to tip the points value up by 10, to having to find 10, and if you do, having no need to look any further
The large events (well, I'm thinking London in particular) often have a lot of inexperienced players, so turning up there and finishing in the middle means you can play fairly averagely but outscore a winner somewhere else.

Also, while I agree that organisers should promote their events, you shouldn't be punishing the players who go to those events if the organisers don't promote them well! You're punishing the wring people! Also, if we're looking at incentives, the points system actually encourages players to not really bother with the events that aren't already big. It's likely to keep the big events big and the small events small.

The stuff I've bolded, I'm not sure I really understand to be honest.

Re: FOCAL scoring system

Posted: Wed Aug 15, 2018 1:31 pm
by Mark Deeks
I feel that with only very rare exceptions, no matter how deep a field is, the winners always go through the same cycle; random draw round one, then two frigging tough follow-up rounds. So to overly weight the points awarded to large fields seems in error, for you aren't playing the vast majority of those people anyway. With tons of respect to Oli Garner, it's not really any more difficult to win a 70+ person event than it is a 30+ person event. Your paths to the final are pretty much the same.

Re: FOCAL scoring system

Posted: Sun Aug 19, 2018 11:48 am
by Graeme Cole
Zarte Siempre wrote: Mon Aug 13, 2018 8:09 am That the mechanic was simple to explain to anyone who happened to ask, to the point where they could reasonably work out exactly how they'd done on the day. And equally that it would be easy to follow in terms of looking over the spreadsheet. For this reason, things with a curve (except perhaps a consistent pre-defined one like the trial we did using... think it was some sort of motor racing scoring) were definitely out.
What if the number of points scored for each position stayed the same as it is now, but there were additionally a fixed number of points awarded to every player for attending an event - say 40? The last-placed player at an event would get 42 points, the second-to-last player 44, and so on, until the first-placed player, who gets 40 + 2 * (number of players).

It wouldn't remove the perhaps disproportionate advantage you get from finishing mid-table at a large event compared to winning a smaller one, but it would go some way towards mitigating the problem where if you miss a large event, it's very difficult to catch up even if you perform very well at multiple smaller events.

For example, suppose person A attends two 20-player events and wins them both, and person B attends one 60-player event and comes 21st. Under the present system they would both score 80, but if each event gave a fixed 40 points to each player, person A would score 160 and person B would score 120.

Obviously 40 is just an example, and this per-event bonus would need to be carefully chosen. If it's too low then it wouldn't have enough effect. If it's too high then the number of events you attended would become more important than how well you played at them.

I don't know if this was one of the systems you trialled, but if so, I'd be interested to hear how it performed.

Re: FOCAL scoring system

Posted: Sun Aug 19, 2018 1:48 pm
by Gavin Chipper
Graeme Cole wrote: Sun Aug 19, 2018 11:48 am What if...

It wouldn't remove the perhaps disproportionate advantage you get from finishing mid-table at a large event compared to winning a smaller one
Then why even invent such a weird system? There are systems that sort that problem out with adding weird arbitrary numbers on the end as an afterthought!

Re: FOCAL scoring system

Posted: Sun Aug 19, 2018 4:12 pm
by Graeme Cole
Gavin Chipper wrote: Sun Aug 19, 2018 1:48 pm
Graeme Cole wrote: Sun Aug 19, 2018 11:48 am What if...

It wouldn't remove the perhaps disproportionate advantage you get from finishing mid-table at a large event compared to winning a smaller one
Then why even invent such a weird system? There are systems that sort that problem out with adding weird arbitrary numbers on the end as an afterthought!
I give an example of what problem it addresses - someone being disproportionately penalised for missing a large event when they did well at a number of smaller ones. What system that sorts the problem out are you thinking of? And does it use a curve?

Re: FOCAL scoring system

Posted: Mon Aug 20, 2018 9:09 am
by Gavin Chipper
Graeme Cole wrote: Sun Aug 19, 2018 4:12 pm
Gavin Chipper wrote: Sun Aug 19, 2018 1:48 pm
Graeme Cole wrote: Sun Aug 19, 2018 11:48 am What if...

It wouldn't remove the perhaps disproportionate advantage you get from finishing mid-table at a large event compared to winning a smaller one
Then why even invent such a weird system? There are systems that sort that problem out with adding weird arbitrary numbers on the end as an afterthought!
I give an example of what problem it addresses - someone being disproportionately penalised for missing a large event when they did well at a number of smaller ones. What system that sorts the problem out are you thinking of? And does it use a curve?
I've already given a system that I think would work. Basically, in a one-person event, they'd get 50; for two it would be 66.7, 33.3; for three it would be 75, 50, 25 etc. There's a formula out there somewhere but I can't be bothered to find/work it out again right now. I think it's the fairest linear system, and the basic principle is that you get points for the proportion of people that you beat, not the absolute number. I think the only downside is that you get non-integer scores, but I can't see why that actually matters really.

Chris Marshall came up with a similar system where the winner gets a fixed 100 points and last place gets 1 (I think) and the rest are equidistant in between. Obviously the winner of small tournaments is then disproportionately awarded and the loser disproportionately punished, but in practice I think the two systems would give a similar-looking table at the end of the season.

Re: FOCAL scoring system

Posted: Mon Aug 20, 2018 9:40 am
by Thomas Cappleman
Gavin Chipper wrote: Mon Aug 20, 2018 9:09 am I've already given a system that I think would work. Basically, in a one-person event, they'd get 50; for two it would be 66.7, 33.3; for three it would be 75, 50, 25 etc. There's a formula out there somewhere but I can't be bothered to find/work it out again right now. I think it's the fairest linear system, and the basic principle is that you get points for the proportion of people that you beat, not the absolute number. I think the only downside is that you get non-integer scores, but I can't see why that actually matters really.

Chris Marshall came up with a similar system where the winner gets a fixed 100 points and last place gets 1 (I think) and the rest are equidistant in between. Obviously the winner of small tournaments is then disproportionately awarded and the loser disproportionately punished, but in practice I think the two systems would give a similar-looking table at the end of the season.
In practice for all but the smallest events these 2 are basically indistinguishable. For example, in a 19 person event the winner under your system gets 95, and Chris' gets 100, and that difference gets smaller as you approach the middle.

I think this is my preferred system (maybe better if you round off the non-integer points). It's still simple enough to describe, and comes closer to rewarding ability fairly. As Mark said, winning a 30 person event is not much easier than winning a 70 person event, but is substantially harder than coming 40th at a 70 person event.

Re: FOCAL scoring system

Posted: Mon Aug 20, 2018 10:43 pm
by Gavin Chipper
The other thing with Chris's system is that the score you get for coming last is arbitrary and a bit like Graeme's idea of awarding points for turning up.

Re: FOCAL scoring system

Posted: Fri Aug 24, 2018 4:50 pm
by Tim Down
Thomas Cappleman wrote: Mon Aug 20, 2018 9:40 am
Gavin Chipper wrote: Mon Aug 20, 2018 9:09 am I've already given a system that I think would work. Basically, in a one-person event, they'd get 50; for two it would be 66.7, 33.3; for three it would be 75, 50, 25 etc. There's a formula out there somewhere but I can't be bothered to find/work it out again right now. I think it's the fairest linear system, and the basic principle is that you get points for the proportion of people that you beat, not the absolute number. I think the only downside is that you get non-integer scores, but I can't see why that actually matters really.

Chris Marshall came up with a similar system where the winner gets a fixed 100 points and last place gets 1 (I think) and the rest are equidistant in between. Obviously the winner of small tournaments is then disproportionately awarded and the loser disproportionately punished, but in practice I think the two systems would give a similar-looking table at the end of the season.
In practice for all but the smallest events these 2 are basically indistinguishable. For example, in a 19 person event the winner under your system gets 95, and Chris' gets 100, and that difference gets smaller as you approach the middle.

I think this is my preferred system (maybe better if you round off the non-integer points). It's still simple enough to describe, and comes closer to rewarding ability fairly. As Mark said, winning a 30 person event is not much easier than winning a 70 person event, but is substantially harder than coming 40th at a 70 person event.
It's my preferred system too and Chris's is the same as mine but with the same value for each event. I only bothered trying to come up with something different because Gevin's had been rejected last time this was discussed. I think we should do rounding for readability, so to reduce the error, I'd suggest increasing the number of points per event to, say, 1000.

Re: FOCAL scoring system

Posted: Fri Aug 24, 2018 8:49 pm
by Gavin Chipper
I agree that it makes sense to do 1000 points if you want to stick to integers.

Re: FOCAL scoring system

Posted: Mon Jun 17, 2019 1:11 pm
by Gavin Chipper
Jeff sent me the Excel file with the 2018 points table so that I could see how things would have looked under a different system.

I used (1 - position / (players +1)) * 1000 rounded to the nearest integer so there would be no fractional points. Basically this means that everyone's score would be between 0 and 1000 for an event, as follows:

1 player: 500
2 players: 667, 333
3 players: 750, 500, 250
4 players: 800, 600, 400, 200

Etc.

It didn't make a massive difference to the top 8. A few people changed places, but only Graeme Cole dropped out, going from 8th all the way down to 12th. Tom Cary took his place going from 10th to 7th. The biggest change within the top 8 was Eddy Byrne going from 3rd to 6th.

But there were big changes elsewhere. Possibly the most noticeable was Edward McCullagh going up from 73rd to 35th. This happened because he won both Newcastle and Dublin (the only events he attended) and these were the two worst-attended events, both with 17 players. The system currently in use gives a big penalty for smaller events. Basically winning one of these two events was the same as coming 51st at London (where there were 67 players) because in all these situations you are finishing 17th from bottom, and that's how the current system measures how well you've done. In a fairly similar manner, Hazel Drury goes up from 63rd to 36th. George Armstrong also goes from 72nd to 46th.

Other players scored highly at London and accordingly drop places. Matthew Tassier goes from 26th to 37th, Oliver Garner (who won CO:LON) goes from 37th to 56th, Conor Travers goes from 46th to 70th, Lauren Hamer goes from 51st to 72nd and Mark Deeks goes from 47th to 73rd.

Currently bigger tournaments are massively over-represented in the points and smaller events massively under-represented.

So let's make the change for next year! :mrgreen:

Re: FOCAL scoring system

Posted: Mon Jun 17, 2019 2:10 pm
by Gavin Chipper
I think one of the best examples highlighting the flaws of the current system is that Josephine Sinclair turned up to just CO:LON and came 30th there, but finished higher in the annual standings than Ed McCullagh who turned up to both Newcastle and Dublin and won both events. It's not even close to a sensible system.

Re: FOCAL scoring system

Posted: Tue Jun 18, 2019 7:33 am
by Matt Morrison
Haha wow. That Sinclair-McCullagh example is brutal.

Re: Countdown in South Yorkshire 2019

Posted: Sun Jun 23, 2019 12:25 pm
by Dave Butt
Hi I am Dave Butt and I am delighted to announce that i will be hosting the 1st ever Countdown in South Yorkshire.

Date: Saturday 17th August 2019
Venue: https://www.premierinn.com/gb/en/hotels ... gKbu_D_BwE
Admission: Free
Hosted by: Dave Butt
Doors open: 10.30am for an 11.15am start
Games to be played: At least 8, plus Grand Final
Finishes by: 7pm
Wheelchair accessible?: Yes…
Dave says… Yes I will be hosting the 1st ever Countdown in South Yorkshire at the Premier Inn hotel in Sheffield it will be in Bristol style 8 10 round heats and 1 15 round final

To enter please reply.

Dave

Re: FOCAL scoring system

Posted: Sun Jun 23, 2019 1:03 pm
by Gavin Chipper
Say hi to Ashton Hancock from me. He lives down your street.

Re: FOCAL scoring system

Posted: Sun Jun 23, 2019 1:06 pm
by Dave Butt
Gavin Chipper wrote: Sun Jun 23, 2019 1:03 pm Say hi to Ashton Hancock from me. He lives down your street.
He doesn’t what are you on about never heard of the man who is he ?

Re: Countdown in South Yorkshire

Posted: Sun Jun 23, 2019 1:35 pm
by Dave Butt
Let’s get back to the main topic.

FOCAL scoring system

Posted: Sun Jun 23, 2019 2:14 pm
by Gavin Chipper
Dave Butt wrote: Sun Jun 23, 2019 1:35 pm Let’s get back to the main topic.
Which is the FOCAL scoring system. Not your Countdown event. Start a new thread!

Re: FOCAL scoring system

Posted: Sun Jun 23, 2019 4:31 pm
by Gavin Chipper
Gavin Chipper wrote: Mon Jun 17, 2019 1:11 pm Jeff sent me the Excel file with the 2018 points table so that I could see how things would have looked under a different system.

I used (1 - position / (players +1)) * 1000 rounded to the nearest integer so there would be no fractional points. Basically this means that everyone's score would be between 0 and 1000 for an event, as follows:

1 player: 500
2 players: 667, 333
3 players: 750, 500, 250
4 players: 800, 600, 400, 200

Etc.
Just another thing about this scoring system - the total number of points available for each event is always 500 * the number of players (subject to some rounding), so 500 will always be the average score for a player, regardless of the number of players in the event. This provides a certain balance meaning things don't get out of hand when two events have massively differing numbers of players.

Under the current system, the average score per player is number of players + 1, so every extra player increases the expected score for every single player involved. That's what has allowed it to spiral out of control to such an extent as the McCullagh example.
Dave Butt wrote: Sun Jun 23, 2019 1:35 pm Let’s get back to the main topic.
Consider it done!

Re: FOCAL scoring system

Posted: Mon Jun 24, 2019 7:36 pm
by Ian Volante
Dave Butt wrote: Sun Jun 23, 2019 12:25 pm Hi I am Dave Butt and I am delighted to announce that i will be hosting the 1st ever Countdown in South Yorkshire.

Date: Saturday 17th August 2019
Venue: https://www.premierinn.com/gb/en/hotels ... gKbu_D_BwE
Admission: Free
Hosted by: Dave Butt
Doors open: 10.30am for an 11.15am start
Games to be played: At least 8, plus Grand Final
Finishes by: 7pm
Wheelchair accessible?: Yes…
Dave says… Yes I will be hosting the 1st ever Countdown in South Yorkshire at the Premier Inn hotel in Sheffield it will be in Bristol style 8 10 round heats and 1 15 round final

To enter please reply.

Dave
There's already a tournament in South Yorkshire in August, it's probably wise to avoid having two very close to each other around the same time, especially as the other one already has some notable attendees.

Re: FOCAL scoring system

Posted: Sun Sep 29, 2019 10:04 pm
by JackHurst
Gavin Chipper wrote: Mon Jun 17, 2019 2:10 pm I think one of the best examples highlighting the flaws of the current system is that Josephine Sinclair turned up to just CO:LON and came 30th there, but finished higher in the annual standings than Ed McCullagh who turned up to both Newcastle and Dublin and won both events. It's not even close to a sensible system.
That's crazy! I'm in strong agreement with the majority of commenters here that the current scoring system needs improving.

In the topic of complexity, somethings like Gevin's system really isn't that much of an overhead to implement or to explain. For the case of implementing it, you literally need to add one column to an excel spreadsheet with a simple formula in it and then drag it down into all of the cells! As for explaining the system to people, just tell them that the more people you beat the more points yet get. If they ask for exact details, tell them to do one (or if you are feeling particularly mean) tell them to go talk to Gevin.

Re: FOCAL scoring system

Posted: Mon Sep 30, 2019 2:56 pm
by Thomas Carey
Gavin Chipper wrote: Mon Jun 17, 2019 1:11 pm
It didn't make a massive difference to the top 8. A few people changed places, but only Graeme Cole dropped out, going from 8th all the way down to 12th. Tom Cary took his place going from 10th to 7th. The biggest change within the top 8 was Eddy Byrne going from 3rd to 6th.
Robbed! (of an E, mostly #don'tdodrugskids #stayinschool #eatyourvegetables)

Re: FOCAL scoring system

Posted: Mon Sep 30, 2019 5:09 pm
by Gavin Chipper
Yeah, not sure what happened with the E.

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 8:06 am
by FOCAL Countdown
So without resorting to a poll -- rather, to keep this well rounded conversation going -- how do we feel about introducing Gev's points system in January 2020?

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 3:30 pm
by L'oisleatch McGraw
lol @ the branded account. Corporate FOCAL has arrived on C4C. :roll:

The "Sinclair-McCullagh example" makes it crystal clear the current scoring system must go. It only serves to make big events bigger and small events smaller (assuming players give a crap about FOCAL points... and we have to assume they do, or what's the point?)

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 3:53 pm
by Zarte Siempre
Whilst I don't mind the scoring system changing if I'm not having to deal with it, I still think Gevin's goes too far in the other direction.

The idea that it's 80% as impressive to win a 4 person tournament as it is a 100 person tournament is just stupid.

If you really wanted to ACTUALLY build a fair system, you'd assign values to the players in attendance, generate an equation based on the strength of the field, and go about it that way.

This is just going "I don't like this imbalance, so I'll throw it the other way instead".

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 4:05 pm
by Thomas Cappleman
Most events though are fairly similar in their spread of strength (with Waterford being the main exception, with a much higher ratio of non-regulars), and also aren't anywhere near 4 or 100 in size. In a small event, there's no way of getting a high position without playing the rest of the top players. In a large event, it's more possible to fluke a high-position, while also possible to finish a lot lower than normal.

I think in general things balance out to fair enough (and MUCH closer than it currently does), so I support the change.

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 4:21 pm
by Jon O'Neill
The change is a vast improvement.

No scoring system is perfect if you want a transparent/easily calculable approach. I can understand why you'd want a transparent system - so you know where you need to finish to get x points on the day, for any event. The Zarte Method is clearly superior to both but doesn't fulfil this requirement.

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 4:28 pm
by Callum Todd
FOCAL Countdown wrote: Wed Oct 02, 2019 8:06 am So without resorting to a poll -- rather, to keep this well rounded conversation going -- how do we feel about introducing Gev's points system in January 2020?
I am in favour.

To those objecting that it isn't perfect: nothing is, but is it at least better than what we currently have? I believe so.

#Gavin2020.

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 6:53 pm
by Zarte Siempre
Callum Todd wrote: Wed Oct 02, 2019 4:28 pm To those objecting that it isn't perfect: nothing is, but is it at least better than what we currently have? I believe so.
Or, we could wait and try and come up with something better rather than going "Ah... whatever, yeah, that'll do"

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 8:08 pm
by Callum Todd
You could equally apply "going "Ah... whatever, yeah, that'll do" to not changing the system yet. If the proposed system is better than the current one (which I believe it is) then why not implement it? It is currently the best system anyone has suggested*. Why choose an inferior* system, just because the proposed system isn't perfect (which no system ever will be, because nothing ever is). How much "better" would a proposed system have to be before we stop waiting?

And, given that the current scoring system has been in place for a few years now and its flaws have been noticed since year one, and still none of the maths/statistics/computer boffins in this community have come up with anything better yet, how realistic is it to believe that "something better" is likely to emerge soon (or even ever)? How long are we going to "wait and try and come up with something better" for? Why can't we just implement the best system currently proposed now, and continue to "try and come up with something better", just as we would have done anyway, while still reaping the benefits of an improved* system in the mean time?

* = in my opinion the proposed system is better. My argument is that, if the proposed system is better (but not perfect) it should still be implemented. Obviously if you disagree that the proposed system is better than the current one, my argument for implementing it now doesn't apply to you. Go argue with Gavin instead.

Re: FOCAL scoring system

Posted: Wed Oct 02, 2019 9:56 pm
by Fiona T
Fairly objectively as someone who has no ambition beyond average to middling, the gevin scoring seems a lot better.

It rewards players who both do well and turn up. Do both, and you're in the finals.

(Although I do think on the parkrun side it should apply to age grading, not finish position!)

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 10:29 am
by Gavin Chipper
Well this is quite an interesting discussion. And although it's being billed as "my" system, I think some credit at least has to go to Tim Down for suggesting sensible improvements - increasing the max from 100 to 1000 so that we can then just round the scores and not bother with decimal points.
Zarte Siempre wrote: Wed Oct 02, 2019 3:53 pm Whilst I don't mind the scoring system changing if I'm not having to deal with it, I still think Gevin's goes too far in the other direction.

The idea that it's 80% as impressive to win a 4 person tournament as it is a 100 person tournament is just stupid.
I don't see why it is stupid. Assuming we want a system that distributes the points linearly and don't want some arbitrary shift (adding or taking away a constant from everyone's score) then this seems to me the most mathematically sound way of doing it.

There are a couple of other considerations as well - such as whether the players at each event are a random sample from the total population of players (basically that if you pick a random player from the 4-player event, they are just as likely to be good as a random player from a 100-player event). But with these assumptions, it is a neutral system regarding the number of players. Whereas the current system favours playing in a large event, the system I've proposed doesn't favour big or small events, so it doesn't go in the other direction. It's dead centre.

So while it might seem intuitively stupid that it's 80% as impressive to with a 4-player event as a 100-player event, this isn't the result of an intrinsic bias in the system. Under the current system it's 80% as impressive to come 20th out of 100 as it is to come 1st. Is that stupid as well? Which is better - 1st out of 4 or 20th out of 100? Under the proposed system, they're about the same.
If you really wanted to ACTUALLY build a fair system, you'd assign values to the players in attendance, generate an equation based on the strength of the field, and go about it that way.

This is just going "I don't like this imbalance, so I'll throw it the other way instead".
The system I've proposed was designed to be unbiased regarding small/large events. That's all. And I think it does that. It wasn't designed to take into account the quality of the field. You can do that, but it opens a whole new can of worms. It would be less transparent and more complicated (see Jono's post), and there would always be some dispute over how the rating of the players are calculated. Would it be based purely on FOCAL events? Would Apterous count? Would TV performances count? How much would you take recency into account? If someone like Conor Travers disappears for three years and comes back, what does that do to his rating? Because we can still be pretty sure he's automatically going to be a contender for the top places.

You could also argue that with smaller events, then even with a random sample from the population, the quality of the field would be more subject to random variance. But that can go both ways, and I think it would be a bit arbitrary to devalue such events on the basis that the field might be weaker than average, partly because it's just as likely to be stronger than average, and it's also an extra complication, and also most events are going to have enough players to render this negligible anyway.

The bottom line is that the proposed system addresses the imbalance in the previous system without throwing it the other way. It is neutral regarding the size of the field. It doesn't address the issue of the quality of the field, but that's a separate issue, and on balance I think it would be unwise to attempt to do that anyway.

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 10:30 am
by Gavin Chipper
Fiona T wrote: Wed Oct 02, 2019 9:56 pm (Although I do think on the parkrun side it should apply to age grading, not finish position!)
Maybe we could run two systems alongside each other!

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 3:28 pm
by Tim Down
FOCAL Countdown wrote: Wed Oct 02, 2019 8:06 am So without resorting to a poll -- rather, to keep this well rounded conversation going -- how do we feel about introducing Gev's points system in January 2020?
Unsurprisingly, I'm strongly in favour.

Publishing the details of the scoring system somewhere easily findable would be good. I've tried and failed to find details of this year's scoring system (specifically, how many tournaments count towards each player's final FOCAL points total. George's report on CO:Liv on the FOCAL site alludes to it being 8 but I haven't seen that anywhere else).

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 5:25 pm
by George Armstrong
Tim Down wrote: Thu Oct 03, 2019 3:28 pm Publishing the details of the scoring system somewhere easily findable would be good. I've tried and failed to find details of this year's scoring system (specifically, how many tournaments count towards each player's final FOCAL points total. George's report on CO:Liv on the FOCAL site alludes to it being 8 but I haven't seen that anywhere else).
I've had a look at the CO:Liv report. For those interested, it reads:

"No change to the Top 8 after Liverpool. Callum has now amassed a clear 116-point lead at the top of the FOCAL table, but, he has also now registered eight events for the year. So, appearances at Blackpool and/or Braintree before Finals Day can only serve to throw out any lower scores from previous events, giving the rest of the field the opportunity to close in.".

Oh, and by the way, Jeff wrote that bit... ;)

But if it is indeed the case that only your best 8 performances then I didn't know that. To think, for an EXAMPLE, if Callum plays both the remaining two events before the finals, there is a chance if he scored enough points at both Blackpool and Braintree (which given the number of expected attendees is unlikely) that the points he accrued from coming 2nd in Dublin and 1st in York would not count towards his total for finals qualification, and that is where the problem with the current system to me lies - IMO finishing highly at any event, regardless of size (let's be honest, the field of players will always be roughly the same overall quality, bar Waterford, so there isn't really an argument there), needs to be properly rewarded, which it clearly isn't at the moment.

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 6:08 pm
by Zarte Siempre
George Armstrong wrote: Thu Oct 03, 2019 5:25 pm But if it is indeed the case that only your best 8 performances then I didn't know that.
It's the nearest value to 2-thirds of the tournaments in a given year. It's just to stop people being priced out of being able to make finals by the people who'd literally go to a co-event inside a wardrobe.

Re: FOCAL scoring system

Posted: Thu Oct 03, 2019 7:25 pm
by Graeme Cole
Zarte Siempre wrote: Thu Oct 03, 2019 6:08 pm
George Armstrong wrote: Thu Oct 03, 2019 5:25 pm But if it is indeed the case that only your best 8 performances then I didn't know that.
It's the nearest value to 2-thirds of the tournaments in a given year.
Which this year is 9, there being 14 events.

(If I've miscounted, please let me know and I will refocalise the Focaliser.)

Re: FOCAL scoring system

Posted: Fri Oct 04, 2019 11:47 am
by Jon O'Neill
Zarte Siempre wrote: Thu Oct 03, 2019 6:08 pm people who'd literally go to a co-event inside a wardrobe.
You get 50 points for winning that event in Gev's system.
Just saying.

Re: FOCAL scoring system

Posted: Sat Oct 12, 2019 12:04 pm
by FOCAL Countdown
Thanks for the input. We'll make the switch after the Wolvo Finals so the two Lincolns in January 2020 will be the first to use the new method.

As per Gev's point, the top end of the scoring range for an event will be 1000, out of preference for integers. And to answer Tim's point, a player's best results from eight events make up their total score for the year. (As an aside, I wrote and hosted a local quiz earlier in the year where questions were scored out of 10, rather than 1, to avoid "give them half a mark" etc and this also made multiple part answer questions more interesting!).

Re: FOCAL scoring system

Posted: Sat Oct 12, 2019 12:48 pm
by George Armstrong
FOCAL Countdown wrote: Sat Oct 12, 2019 12:04 pm And to answer Tim's point, a player's best results from eight events make up their total score for the year.
To clarify, does that mean that your top 8 performances pointswise for THIS year count towards your FOCAL points for the Finals?

Re: FOCAL scoring system

Posted: Sat Oct 12, 2019 2:26 pm
by FOCAL Countdown
That is how it works, sum of the points in up to eight events. Highest finishing position at a single event would then be used in the event of a tie (this appears as a column on the league table).

Re: FOCAL scoring system

Posted: Thu Sep 15, 2022 7:23 pm
by Jack Neal
Why are 2020-22 events capped at 8? The rules as in this thread viewtopic.php?p=165388&hilit=focal#p165388 state that it is capped at 66.6% which this $eason would be 13.

Re: FOCAL scoring system

Posted: Thu Sep 15, 2022 7:45 pm
by George Armstrong
If we changed it to 13, the top 8 would in theory be more weighted to those who have done more events rather than those who've performed well. As it stands, 1 person has played 13, with a further 4 likely to. Only 1 person who isn't in the current top 8 would be if we'd extended it to 13, and they've still got a chance of making the finals anyway. Ultimately it doesn't matter that much.

Re: FOCAL scoring system

Posted: Thu Sep 15, 2022 9:21 pm
by Jack Neal
George Armstrong wrote: Thu Sep 15, 2022 7:45 pm If we changed it to 13, the top 8 would in theory be more weighted to those who have done more events rather than those who've performed well. As it stands, 1 person has played 13, with a further 4 likely to. Only 1 person who isn't in the current top 8 would be if we'd extended it to 13, and they've still got a chance of making the finals anyway. Ultimately it doesn't matter that much.
Why don't you do it the average of all events up to 13. So the formula is Total Points/Events Played *13?

Re: FOCAL scoring system

Posted: Sat Sep 17, 2022 2:47 pm
by JackHurst
Jack Neal wrote: Thu Sep 15, 2022 9:21 pm
George Armstrong wrote: Thu Sep 15, 2022 7:45 pm If we changed it to 13, the top 8 would in theory be more weighted to those who have done more events rather than those who've performed well. As it stands, 1 person has played 13, with a further 4 likely to. Only 1 person who isn't in the current top 8 would be if we'd extended it to 13, and they've still got a chance of making the finals anyway. Ultimately it doesn't matter that much.
Why don't you do it the average of all events up to 13. So the formula is Total Points/Events Played *13?
Because this rewards too much somebody who goes to one event and wins it but them goes to no others.

On the flip side, Imagine you wanted to qualify for focal finals really badly, and you won the first event of the season and it was attended by 60 people. After this point you'd effectively get penalized in the rankings for any further events you attend...

The current system is good :)

Re: FOCAL scoring system

Posted: Sat Sep 17, 2022 2:54 pm
by Gavin Chipper
George Armstrong wrote: Thu Sep 15, 2022 7:45 pm If we changed it to 13, the top 8 would in theory be more weighted to those who have done more events rather than those who've performed well. As it stands, 1 person has played 13, with a further 4 likely to. Only 1 person who isn't in the current top 8 would be if we'd extended it to 13, and they've still got a chance of making the finals anyway. Ultimately it doesn't matter that much.
So will there have been 20 events in this championship? 8 out of 20 (40%) seems quite low, but I suppose it's about finding the right number for the job at the time. With the drawn-out championship over three years because of the pandemic in the middle, it's more understandable that people are less likely to make 2/3 of the tournaments. Do you think it will be more like 2/3 in future? Or do you think that's just generally too high now?

Re: FOCAL scoring system

Posted: Sat Sep 17, 2022 8:47 pm
by Jack Neal
JackHurst wrote: Sat Sep 17, 2022 2:47 pm
Jack Neal wrote: Thu Sep 15, 2022 9:21 pm
George Armstrong wrote: Thu Sep 15, 2022 7:45 pm If we changed it to 13, the top 8 would in theory be more weighted to those who have done more events rather than those who've performed well. As it stands, 1 person has played 13, with a further 4 likely to. Only 1 person who isn't in the current top 8 would be if we'd extended it to 13, and they've still got a chance of making the finals anyway. Ultimately it doesn't matter that much.
Why don't you do it the average of all events up to 13. So the formula is Total Points/Events Played *13?
Because this rewards too much somebody who goes to one event and wins it but them goes to no others.

On the flip side, Imagine you wanted to qualify for focal finals really badly, and you won the first event of the season and it was attended by 60 people. After this point you'd effectively get penalized in the rankings for any further events you attend...

The current system is good :)
Hi Jack,

I meant only use the formula for people who attended more than 2/3 of events. So your scenario wouldn't work.

Jack

Re: FOCAL scoring system

Posted: Sun Sep 18, 2022 2:54 pm
by Matt Morrison
Can you two please keep it civil, otherwise we'll all be forced to witness you Jack-off.

Re: FOCAL scoring system

Posted: Sun Sep 18, 2022 4:14 pm
by JackHurst
Matt Morrison wrote: Sun Sep 18, 2022 2:54 pm Can you two please keep it civil, otherwise we'll all be forced to witness you Jack-off.
Jack it in man.

Re: FOCAL scoring system

Posted: Mon Sep 19, 2022 4:58 pm
by Jack Neal
JackHurst wrote: Sun Sep 18, 2022 4:14 pm
Matt Morrison wrote: Sun Sep 18, 2022 2:54 pm Can you two please keep it civil, otherwise we'll all be forced to witness you Jack-off.
Jack it in man.
Very funny. I'm not even gay!