Slitherine

The new FoG Rankings 2008 section is now available in the "on the field of glory"

It appears that the rankings are only for British players. Is there any thought of putting up rankings for other countries? (I can certainly understand that there may currently be sample size issues with ranking for other countries at the present time.)

Also, what are the criteria used to determine the rankings?

Marc

No, there is some frenchies (Jérôme Bodelle)
But It's true it's only for the british tournaments ? And same question : what are the criteria used to determine the rankings?

These rankings are actually the BHGS rankings, I maintain them and am looking at adding overseas events but need to decide how to do it.

The BHGS rankings system is a 12 month rolling system that calculates your rank based on your possition in all the comps you play during the year. Comps are rated Minor, Major and Grand Slam. Minor events are worth 80 points, Majors are worth 100 and Grand Slams 110 althought these values may change. Winning a comp gets you the full points, finishing last gets you 1 point and placing in between gets points on a sliding scale. These points are added up and then divided by 6 to get a rating out of 100. If you play more than 6 events Grand Slams and Majors count as a full 1/6th each and any minors are averaged out to make up the remaining places, there were 3 grand slam events, 2 majors and lots of minors.

It would be perfectly possible to use exactly the same system to do a unified FoG rating system to get an idea of who is doing well internationally.

This is most definitly not Glicko.

That is an interesting system. Do you find that there is more movement possible than with Glicko? And how do you decide whether an event is minor, major, or grand slam? Is it based on the number of players involved, or perhaps the perceived "importance" of the event?

Curiously,

Marc

It is the historic BHGS system and definitely has a lot more movement than Glicko. It may be that in the UK Glicko moves so slowly because we played so many games but essentially Glicko was tried and almost universally despised over here. The BHGS rankings have been the topic of many a conversation at clubs and you get lots of rivalry even down in the lower reaches of the rankings.

I can remember getting excited when I first made the top 100, getting into the top 50 felt like a real acheivement, when I hit the top 20 I was thrilled. I have managed to get to number 1 in the UK DBM rankings a couple of times but never managed to hold it together for a full calandar year It certainly gets people looking to enter comps.

The classifications are to some extent based on importance and to some extent based on history.

At the moment the Grand Slams are Britcon and the BHGS Challenge.
The Majors are Warfare and Roll Call, Derby was a major but is going to drop to a minor as there really were not many players there last year.
The Minors are everything else including all the doubles.

I can very easily add in every foreign tournament to the mix, it might need a bit of jugling if we make lots of other Grand Slams and someone enters too many Grand Slam or Major events.

I am tempted to put the WIC in as a major or grand slam event and am open to suggestions for which other overseas events should be considered other then minor. Historicon for example should be a grand slam.

hammy wrote:
This is most definitly not Glicko.

And makes a great argument for Glicko. As its all tosh.

There is indeed more movement than with Gecko - at least once gecko has a significant amount of data entered into it, I would think initially that'll have a lot of movement as well. Certainly when we tried gecko for DBM in GB - and a lot of data was fed into it to set it up, years of results - the players didn't like because of the lack of movement.

Classification of events is, more or less, on perceived importance - and singles events are always more important than doubles ones.

Not scientific but a bit of fun designed to encourage participation at comps - hence you cannot maximise your score without attending at least 6 comps.

We have been using glicko system in Spain for many years, but it´s not a popular system, in 2008 we are using a champions race, but we do not have a definitive ranking system. I think it´s time to do a "Papal Conclave", and designing a ranking system, at least it can be implemented in the countries of the FOG Zone.


Hammy maybe the IWF has something to say about it?


Regards

David

robertthebruce wrote:

Hammy maybe the IWF has something to say about it?

No doubt eventually, but I wouldn't hold my breath

I suspect that the IWF will want to do some sort of Glicko type system. Personally I have lost all interest in Glicko because while it is scientific and with all the work done regarding international weightings produces IMO quite sensible rankings it doesn't produce much in the way of player interest.

The BHGS style system does make the race each year a lot more fun and the years champion is often undecided with just the last event to go.

Toby Partridge who is running the DBMM rankings is using a system based on the BHGS system but has written some nice web based software that will allow for things like looking at player by player stats or army by army stats while still using the system I described before.

It would be great if we could get an international FoG ranking system, I really don't know what is the best way to do it. So far there have not been that many international games, just the BHGS Challenge really. There will be some more in Helsinki and I would expect some UK players to be going to France before too long and I know that several French players are looking at coming to Britcon.

Perhaps if we had international rankings where each country got to nominate one event as a grand slam, any FoG event at the IWF World Champiionships should also really be a grand slam. Singles events other than the nominated grand slam with more than a certain number of players should be majors, other comps minors or something like that.

I would go with a much more dynamic system than Glicko and lets get it up and running, so we can have some fun.

Remember we also have the ladder system available on the site. We could just dump all the results into that?

If we go with an international ranking Hammy is happy to do the book keeping so all we need is the major and minor events from the US, France, Italy, Spain etc

Slitherine will present a prize for the winner but it must be inclusive of all of the territories that play the rules.

So lets get organised

JDM

hammy wrote:I suspect that the IWF will want to do some sort of Glicko type system. Personally I have lost all interest in Glicko because while it is scientific and with all the work done regarding international weightings produces IMO quite sensible rankings it doesn't produce much in the way of player interest.

An excellent description of Glicko: "sensible, yet boring."

hammy wrote:Perhaps if we had international rankings where each country got to nominate one event as a grand slam, any FoG event at the IWF World Champiionships should also really be a grand slam. Singles events other than the nominated grand slam with more than a certain number of players should be majors, other comps minors or something like that.

Now that's an interesting idea. And best of all, JD has volunteered you to do the work. As an aside, the Glicko system does have the real benefit of capturing a lot of data that can be put together in interesting ways. In fact, a lot of the appeal of Glicko here in the US has been the oddball stats such as player nemesis ratings, most armies used, best score with army, etc. Whatever system ends up getting used (by Hammy ), it should include a way to track all these bits of side data.

Marc

Exactly marc the glicko other stats were far more interesting. Furthermore there was a ton of movement in the player of the year results. But that was never observed by how glicko was done in the UK.

But either way, I strongly suspect there will be no glicko in the FoG world as the Fog-ites and glicko-programer types are not in sync.

So we do need to move on. I do think something that attemtps to define best is not the ideal goal, but something that tracks and reports usage and frequency and all that is more likley to have enduring interest.

What may be worth doing would be to take advantage of the code written by Toby Partridge, that has additional features like army by army results and such.

The DBMM rankings at http://www.dbmm.org.uk/rankings/index.php are done using Toby's code and you can to some degree drill down by player or army.

Toby has offered his software and I might even be able to get him to add an international facility too.

I'm very much in favour of a model based on the BHGS one described above. To encourage game playing the lowest strata of points value could be club games (with the second strata being 'any other con' and the highest strata annual National and World titles) this would mean that even in countries with small populations FoG players can still work hard to build up their points towards an international score. It's important to break out of regional thinking when designing a leaderboard for an international game, there's nothing wrong with allowing users to filter data by country and having prizes for regional champions - but a global score is absolutely essential. Those familiar with Magic the Gathering will understand (our system has simpler and faster calculations but the principle of having a global perspective that can be filtered by region is the same). A mess of uncomparable French/British/Norwegian ladders is not desirable - and this is our opportunity to dodge that ball!

Implementation could be as simple as taking Toby's code and hosting it on the Slitherine site; anyone can look at the data but only regional reps (trusted volunteers) can submit results.

Now of course there's the moaning of "who's going to run it?" but the answer is that when it is correctly deployed the users will run it. The top of the heirachy makes some hard decisions about who the regional reps will be and chases them for results but otherwise the system runs itself. Head judge brings a laptop to the world title, submits the data, and everyone around the world can see who this year's winner is within a few seconds!

~Aside~
We seem to have lots of people that have purchased FoG in Australia but they don't seem to be coming to cons yet (hence my club-level-points flag waving!). My humble opinion is that within a year we will have enough FoGgers that have migrated from DBx to start getting meaningful data from cons. Sadly the inaugural GenConOz will not have a FoG event ~ cross fingers for next year

**** edited ****

The Glicko stats for DBM were fascinating.

**** edited ****


ADMIN

We will not accept bad language on the forums. Take this as a warning.

Rest of message edited to remove offensive material.

OK, including club level games is a nice idea but the BHGS style rankings system doesn't really support individual games. It is designed to work on how players fare in tournaments or at least a series of games. Club competitions or ladders could be included but I can't think how you would include single games.

Glicko does let individual games count and you could have any reported game counted for rankings. Where Glicko has problems is that if you have several independent player pools or pools with limited crossover it does not cater for differences between the relative strength of the pools. For the international DBM Glicko to get to where it is now has taken a lot of work and a lot of international games and there is still a significant amount of fudging to get what seems a reasonable set of results. The problem is that if you have a club where one player wins every game he playes but nobody at that club plays outside the club then Glicko would treat that player as the best player in the world while he may just be an average player with a lot of very weak clubmates.

There is a ladder system on the Slitherine site, I suspect that for club games the ladder may be the best place to record things for now.

If someone wants to stand up and run Glicko for FoG then nobody is going to stop them. Unfortunately Glicko has a very bad press in the UK simply because (all the following are my personal feelings about Glicko):
* it is intensly boring
* the vast majority of players had ratings that hardly ever changed
* there were several semi independent playing pools and the big fish in a small pond syndrome was rather evident
* players reliability almost never got worse despite hardly playing (in UK terms)
* once at the top of the rankings just playing once a year was more than enough to stay there
* the UK software didn't have much if any of the more interesting info that other countries had

I accept the BHGS style rankings are not 'scientific' and that to get a good ranking for the year you have to play in six events but the main reason for the rankings is to get people to play games, not to get a great Glicko rating then sit on the sidelines and bask in the glory.

*MODERATOR COMMENT*
Phaze_of_the_Moon,

If you want to comment on this thread that is fine but can you refrain from crudity, I am sure you are more than capable of putting your point without it.

hammy wrote:I suspect that the IWF will want to do some sort of Glicko type system. Personally I have lost all interest in Glicko because while it is scientific and with all the work done regarding international weightings produces IMO quite sensible rankings it doesn't produce much in the way of player interest.

Politely, that's because whoever implemented it in the UK frankly made a right pig's ear of it. All you had was the "dull as watching
paint dry" long term rankings as far as I could tell. I'm not surprised nobody was interested in it. Railway timetables are
more interesting. As Dan and Marc point out, the player of the year race is as variable as the current BHGS ranking system,
plus you get all those stats on player's arch-nemesis, who-played-who, army results, etc etc.

Rgds,
Peter

Actually all those are quite dull as well after the initial interest wears off - much more fun is wild speculation based on unscientific opinion that leads to trash talking and grudge matches arranged over too many beeers