Talk Fantasy Football

***New global glicko 21-Jan-20 - see this post.***

Hi all,

Some exciting news for NAF members today, as we are launching our new Glicko ranking system, implemented by Nick Harding / mubo.

These rankings aren't designed to replace the Elo numbers you've been used to since NAF rankings began, they are simply a new way of processing NAF data and ranking coaches, and will run in parallel. The Glicko system is able to introduce new features that some coaches have wanted for many years, such as decay and greater reactivity. You can read more about how they work here and view the ranking page at thenaf.net here.

The system is currently in a 'soft launch' period, and if you experience any difficulty viewing the page or using the dropdown filters, please flag this up. The rankings will update on a monthly basis, and we hope the April update will be take us to a fully 'live' state. As we move forward, you can expect fun and interesting updates and developments to pop up on both thenaf.net and thetacklezone.net, and we'll advertise these as and when they become available.

Finally, we'd like to acknowledge Kyrre for his great work parsing the games played in the database so Nick could work with the data and implement the system.

Please join me in virtually patting Nick on the back for his fantastic work, and enjoy playing with the new system!

Thanks,

Phil.

Seems a decent system Noticed that 2 of my races have been artificially high in the ELO ranking (Nurgle 2 tourneys, Pact 1 tourney) as they happened to get decent results on short runs.

Good work to those putting the effort in.

Here here!!

What a time to be alive!

Kudos to mubo, clearly that man has got it all Complimentary beer at NAFC coming up!

Great work mubo!

straume wrote:What a time to be alive!

Quite!

Really interesting stuff! Thanks NAF, Mubo and everyone involved.

Couple questions - does the size of the tournament, or whether it's listed as a Major event have any bearing on Glicko scores?

Secondly, is the longer term aim to replace the current ELO based stats with these? Ultimately we want one official set of rankings surely? So if these are felt to be an improvement would we adopt them in full at some stage?

Morning, Brendan. The answer to both questions is no.

On the first point, the system cares not for the size or gravitas of the tournament you played at, just the game of BB you played and who you played against. On the second, it’s not our intent to reduce the choice available to members or take away something some people have loved for years. You can choose to follow both systems, one system or neither, it’s entirely your choice depending on what you like!

We need two world #1's

More things for people to argue over

Purplegoo wrote:Morning, Brendan. The answer to both questions is no.

On the first point, the system cares not for the size or gravitas of the tournament you played at, just the game of BB you played and who you played against. On the second, it’s not our intent to reduce the choice available to members or take away something some people have loved for years. You can choose to follow both systems, one system or neither, it’s entirely your choice depending on what you like!

It is very early days of course but I do wonder whether having two ranking systems, evidently with equal status from what you write above Phil, is going to be a good thing. My observation is that having official rankings is a real strength for the NAF; coaches love monitoring their progress, checking others’ progress. It’s something which drives the tournament scene beyond just the trophies and prizes on offer at events. Two systems for members to compare or to decide which they prefer seems confusing.

It runs a risk of mirroring the Bret and Khorne ruling i.e. where NAF acted on the membership’s democratic choice to include them, however decided that TOs can choose to disregard them if they wish. I would have preferred a strong NAF decision i.e. B&K are either official (like the other 24) or not. For me a clear decision, at least after a trial period, would be good regarding which statistical model is the official NAF ranking.

Rankings are a silly bit of fun, right? People will interpret them whichever way they want. But if the new ones are basically an improvement, then why keep two systems and confuse everyone? Or indeed if the Glicko numbers are just an alternative exercise with the ELO system remaining the official rankings, then clarity on that would be helpful too.

Don’t get me wrong, my overwhelming observation is that this a fantastic new feature and I applaud all the work done, especially Phil as TD for acting on comments and implementing this. I Just feel in the longer term it would be good to work towards a single official ranking, rather than a choice between different number crunching models.

I appreciate all flavours of feedback, so thanks for posting again Bren.

I agree that rankings are nice things to have and a strength for the NAF; some coaches really follow the numbers and I think most members are at least roughly aware of how they're doing (Elo wise at least, Glicko is yet to even be fully launched). It's great we can offer this sort of tracking. However, I don't agree that having two systems is going to get particularly confusing. If numbers are your thing, more is better, and crucially, choice is good.

The text on thenaf.net keenly points out that we're not presenting Glicko with the claim it’s an improvement to Elo; it does things in a different way and rewards different behaviours. If you are the sort of coach keenly pursuing the 26, you may be more interested in the Elo system because you aren't being temporarily penalised for leaving your better races behind while you chase your patch; you can still compare your performance to others even while you're doing other things. If you are a coach that plays 3-4 races with regularity in prolific environments, the Glicko page might well appeal to you as a reactive, agile snapshot. It's OK to have different methods that work for different people.

For the foreseeable (post 'hard' launch), we'll run with two 'official' ranking systems, and coaches can choose what to follow, if anything. If Glicko organically becomes far more popular than Elo, then that will drive what people consider to be the big one anyway. In some regions the doubleskulls 'all in' ranking is already seen as better / more important than the NAF Elo numbers, so that's not so much increasing confusion as creating preference. Offering another way of looking at it will, I think, drive a similar sort of discussion / choice. All things will be continually reviewed, so this stance may change one day, but for the foreseeable, I think this covers where we'll be.

Thanks for the response. I can see there's a broader picture. Glicko may not be what everyone is after in every region. I still feel it would be positive for NAF at least at some point to retain one set of official NAF rankings, so we can say who is the outright world number one, rather than having two or three number ones.

On your particular point about the alternative "Doubleskulls" rankings being analogous in some way to these - those stats hold no official NAF status, are very infrequently updated, and are not widely known or understood. The Glicko numbers on the other hand are official NAF rankings, presented as an equal alternative to Elo. That seems quite different so the dual system will be much more evident to the membership at large.

Anyway I look forward to following both and seeing how things pan out...

26 races, 2 sets of rankings. We have 52 World Number 1's

Pack of cards as a tourney giveaway?

Plus jokers...

US Military is way ahead of you...