We're planning on adding a 'rankings' section to the agotny.net web site. We hope to have separate rankings for: LCG joust, standard joust, and multiplayer formats, as well as individual win/loss records of a player's chosen house.

For everyone who wants to join, please email me ([email protected]) or post under this topic the board names or email addresses of the players from your tournaments (each tourney must have at least 4 players), the results of each match, the players' respective houses and the tournament format used. We will accept results from tournaments taking place from Jan 1, 2009 onwards.

Also, if anyone can forward a decent ELO ranking program or link it would be appreciated.

Comments and suggestions are welcome.

oh sweet! Nice show of initiative guys! also, my meta doesn't really use email or the boards, other than myself anyway. is that a problem?

Just first name or nickname then and the meta where they are from. These rankings will naturally depend heavily on the honor system but independently verifiable board name/email address is always nice.

ahhhh good call. ~maybe if we could text our results in. haha

http://www.elo-ajedrez.com/Calculadora_elo_m.php

Try this online Elo calcularor. It's easy and great!

player rankings tab is up on agotny.net , at this point we only have data from the ny meta

Everyone, get your tourney results in!

we'll have one for you guys on thursday.

I'm digging up results we've had in tourny type settings (i.e., League nights and Hedge tournament) going back to the start of the League's "first season" and passing them along to Sith little by little, so if you notice over the next few days that the NY players have many more reported games under their belts then what is shown right now it's not because we're having a monster 24hour AGoT session right now. (Ah, if only.... )

check your g-mail!!!

Sithlord, correct me if I'm wrong, but I believe you need the following info for reported matches:

For joust: Date, Round #: Player A (playing house) defeated Player B (playing house), deck format (i.e., LCG (including League) or standard)

If it's a melee: Date, Round # (if applicable): Player A (house) >[finished ahead]/=[or tied] Player B (house) >/= Player C (house) [and so on]

And then for each player: Player's first name, FFG/agotny "handle", email address, and state (or country if outside US).

Btw, no need to sign in to the agotny.net site, other than if you wish to post to the agotny forum (which we encourage folks to do, of course). Ideally the players shown in reported games would have an online "presence" at FFG or agotny, but I don't think it's required.

yeah, basically to enter ELO results I need who played against whom in each round and the outcome (win, loss, draw). Actually if it's your first match entered I just need your win/loss record because everyone enters with the same score of 1500, but otherwise I need specific match up results.

I'm realizing multiplayer scoring might get trickier, but for now we need the place of each player at each table.

~Wait a minute, so my 250 wins against Staton will still result in a drop in my ELO?!

Maester_LUke said:

Any game played against Staton will result in a drop in your ELO.

JerusalemJones said:

~Because no one ever wins against Staton! Unless your name is Andrew and you are playing GJ.

For what it's worth, there's a *much* better scoring system that is statistically sound and can incorporate joust and melee matches. Drop me a line if you're interested.

is that the one we used at gencon last year?

thebraz said:

sure just post the info/link if available.

Lars said:

Heck no. It is based on Microsoft's True Skill rating/ranking system. They left a couple of open questions regarding decay in skill that I have answers for, but the basic system they've developed is rock solid statistically. The main page for the system is below.

http://research.microsoft.com/en-us/projects/trueskill/default.aspx

If you don't build your own database, you have to enter and update the info by hand using one of their calculators. I've done that for little leagues around here and found that the skill ratings converged quickly to a "realistic" representation of skill e.g. Rick Harmon was miles ahead of the trips and me.

the link only has example calculators. Do you have a link to a calculator that I can actually use?

@thebraz: There's some fascinating stuff on that web site. Thank you for pointing it out.

@sithlord: Did you see the page at research.microsoft.com/en-us/projects/trueskill/calculators.aspx? It has a link at the top to two calculators (labeled "form based" and "AJAX based") that appear to actually crunch numbers for you. I don't know how easy it will be to use for what you want, though. The calculators clearly need some other system for tracking results of each calculation; maybe thebraz has an Excel spreadsheet or something that he could make available.

There is also a fascinating article at www.lifewithalacrity.com/2006/01/ranking_systems.html that discusses several different ranking systems that have been developed in the past few years, which are used for both two-player and multi-player competitive environments. (I found that link, and other interesting ones, by looking through the Microsoft site mentioned above.)

Speaking theoretically, I foresee one complication with using something like TrueSkill rankings for our purposes. Namely, as I understand the system, each meta forms a different population of test subjects, so the variables "mu" and "sigma" might not be comparable to other metas. The top-ranked player in Meta X might have a higher "mu" than the top-ranked player in Meta Y simply because Meta X has less skill depth. If those two players later meet and play at a GenCon or regional, I presume the system will adjust -- but it might take several games to convince the system that the first result wasn't a statistical fluke. And what about all the players from Meta X who never traveled to Gencon or the regional? A perfect system would somehow take the GenCon/regional results as indicative of the overall skill level for Meta X, and adjust accordingly, but I doubt that such a sophisticated system exists.

Thebraz: you apparently have a statistics background. Does what I'm saying make sense, or is the TrueSkill system somehow able to compensate?

Arma virumque said:

The whole theory behind Baysean approximations is that it makes the best approximation possible with the given data. As more data becomes available, the approximation becomes better, but it can't predict based on information it doesn't have. You'll run into similar issues with any other ranking system. At the risk of being pithy: it compensates by updating after every match. When two players who have not played together play and if they have different mu and small sigma and an unlikely result occurs, the sigmas increase and the mu are radically changed because the system recognizes "oh, I didn't know what was going on."

The Excel spread sheet I had for tracking the results after running it through the form based calculator is on my old laptop that needs a new battery, but it isn't too hard to put together. You just need to track mu and sigma for each player.

God - I hate when you guys get all math-y.

Oh, was that math? I wasn't even entirely sure of that any more.

I thought they were talking about frats and frat parties...ah good times...."who's birthday is it tonight?" "Jim's?" "Crap, i knew we should have made it Nezzir's..." True Story (isn't that a staton qoute?).

LOL. Sorry. Here it is in less mathy / fraternity terms:

The ranking system used by Microsoft assigns a Skill Level to every player, which fluctuates up and down as he plays games, but eventually settles within a narrow range. For example, after 30 games, I might be on average a Level 30 AGOT player, and you might be a Level 20 AGOT player.

However, because we are playing in different metas, my Level 30 can't be compared to your Level 20. You may actually be better than me, but because your meta has more highly-skilled players than mine, you don't win as often as I do.

So far, that's pretty obvious intuitively. Here's where it gets interesting: If I travel to your area and you whomp me several games in a row, the system adjusts: For example, I might change into a Level 23, while you might become a Level 25. (Behind the scenes, the system will also adjust another figure that shows the uncertainty of the previous calculations; that's what thebraz was writing about, but I'll ignore that for now.)

However, now my Skill Level can't really be compared to my old meta anymore either. Pretend that before my trip, my brother had a Skill Level of 26. But after the trip, my Level has dropped to 23, while his is unchanged. Now I appear to be a worse player than he is, which would be galling....

Furthermore, your new Level 25 can't really be compared to your old meta anymore either, for the same reason.

With time, if we each continue to play exclusively within our metas, our Skill Levels will return to a new equilibrium. But then they won't be comparable to each other anymore, once again.

Why is this important? It's not, really. As thebraz pointed out, every ranking system has similar problems if there's not a lot of "inter-league play." It's really only important as a warning not to take rankings too seriously -- a warning which I think most people in these forums don't need.

(Oh, yes: It may also be relevant to Sithlord's decision on how much time he wants to spend managing the ranking system. That was actually my original motive for joining the conversation.)