@Chrispminis "I don't see why World is so heavily devalued"

Because if it weren't a WTA win in them would be +a gazillion points. (you can just do a calculation, with 17 equally skilled players you'd gain +91% of your rating by a solo)

"Well, my only issue is that the numbers chosen seem to be quite arbitrary."

True, and so is the number for the variance of the classic diplomacy game. I certainly am not saying that these are definitively "right" or anything, or that I won't change them either.

At some point I'll start looking at the strength of the correlations between G-R and variant results

@ spyman "I am in favour of live games (and I guess gunboat too) not counting for points and not counting for Ghostratings. It would good to have a category of play where players could try out new, possibly riskier strategies etc, without fear of compromising their hard won rankings."

I'd also love to have unranked games.

@TheGhostmaker and spyman - If we are arguing the merits of different kinds of diplomacy, then there is no way that live games can count for less or nothing. Since most of the arguments on this thread seem to be about the purity of diplomacy, live games are the only truly pure form of the game. Remember...15 minute phases.

jimgov I think live games are worthy, and ideally they would sit in a category of their own. Likewise for gunboat and public press. Perhaps a comparison would be the 100 metres in athletics, versus the 800 metres - they are both worthy events, but they are different events.
Unfortunately though, live games seem to be afflicted by more problems than the longer games. I suspect the drop out rate is higher (civil disorders), plus these seem to be the favourite games of the cheats - those who are harvesting points and wins for their profile.
At any rate however things are organized unranked games would be a bonus, whether those be live, gunboat or normal games.

I am not a fan of the ratings system - but that is a technical issue.

I would ask though, if the stats are there, why not compile separate ratings for WTA standard games and then each of the variations...?

Okay, a lot more work, but if one person controls the final output (formulae, publication dates etc), a few jockeys could be 'hired' to do the rest of the work.

Those who love live games should be able to supply a jockey and so on... It would certainly make this site stand out and we could also gauge our opponents better.

Nice idea but I think it only has practical meaning if it's implemented in the profile pages as well so that you can easily see your opponents' rank in various game types.

hellalt +1

Unfortunately, we've been there and Kestas doesn't want to make them official. Seeing as it is his site, it is his call to make as well.

Well I think it's becoming more and more common for people to put their own GR's in their profiles, so it's happening in an unofficial grassroots sense.

TGM, I suppose that's a decent point to make, though I would argue that it is indeed quite a bit more difficult to solo in a world game against 16 of your equals than against 6. At the very least a solo in a world game should have the same effect as a solo in a normal game yes? So maybe you could adjust the percentage such that this will be true?

That would be just as arbitrary. As I say, this kind of thing really needs to be reviewed

Wo bu dong.

Timur? WTF?

I think it's easier to win in world map variant
Since the map is huge if you happen to have a few rookies as neighbors you can grow really fast and then you are unstoppable.
Actually, I consider world map variant a crappy faulty fake copy of original diplomacy and I have seen a rookie with 45% defeat ratio win one of them so I really don't think you need skills.
Just damn luck.

Or a few friends... I fear that has ruined my games. Mind you, the games where everyone is kosher are incredibly hard to win.

I believe Timur's saying "I don't understand" in Mandarin, though my Mandarin is pretty rusty.

hellalt, give me a game where I'm Turkey and have rookie Russia and Austria and I'll give you a solo. Well, a likely solo anyway. I'm in an anonymous world variant game right now, and to be fair, it was started with the 17 highest Ghost rated players that wanted to play, and it's been really fun and engaging and I'm flexing my diplomatic chops more than I have in probably a year. Though, that could just be that the novelty of a variant has spurred me on out of the low effort non-committal slump I've been having due to a declining interest in the game.

I ll give you this one:
gameID=18781
check out the 3 players that drawn the game
1 good player and 2 mediocre ones
in fact one of them has 66% defeat ratio...

In defence of the world variant - there are also higher pot good quality games (just like there are in standard diplomacy). If I need an example - I would choose this one given the standard of players involved: http://www.webdiplomacy.net/board.php?gameID=19317

How is the South Pole a supply centre?

Never underestimate the military applications of penguins...

Here's my honest opinion...

TGM makes his rating system on his own time outside the actual diplomacy rankings. If you don't like them you have multiple options:
1.) Don't follow them, don't read in to them, ignore them
2.) Make your own system, implement it, and get people to follow it. I'm willing to bet TGM would put your rankings right next to his on his own site if you were nice about it, and could show that they are worthy rankings.

So... thanks TGM for all your hard work. I like it how it is.

If you want my real opinion, I think lulzworth had the right opinion up front:

- World and ancient variants aren't as well known and studied as the original, and isn't played nearly as much... plus it is a whole different board. Apples and oranges.
- The fact that there are so many players in world and so many enemies, one cannot diplom fast enough in most cases, and someone can get far too much out of a single game.
- I'm not a fan of gunboat, so I am biased, but the game is diplomacy, so there is no way you can say a variant that doesn't follow the primary form of playing (diplomacy), we are pretty much talking a whole different game.
- Live games, while it looks like a 'real' game as far as time, typing, receiving comments, etc is slower online (unless everyone is in a video chat), and other pieces, like trying to hear conversations and stuff make really not that close to a real game. And the fact that most players bail out before the end makes it not that great of a game.
- Anonymous makes the game fun, since you can't play people you know's strength and weaknesses, and you can play a completely different game than you are used to. But even though most would view it as 'purer' (including me), doesn't that change the game?

The only change I'd like to see (if it isn't in there already) is the ability to not include rankings where someone left the game with more, than, say 3 SCs. Cause that gives one or two players an unfair advantage compared with the others. Thoughts?

Agreed. Especially if they like left the game before Spring '03

I know draws before a certain year aren't counted.

TGM, I know you explain the formula publicly, but have you/will you make the algorithm public? (for the record, I have no issue if you want to keep it private, just curious if you intend on it being public)

I have already. The article I wrote allows you to construct the algorithm, and it is also explained on the dev forum.

Is not most of the algorithm explained here?
http://sites.google.com/site/phpdiplomacytournaments/theghost-ratingslist

It is.

@Fortknox, @Draugner, @TheGhostMaker
What if a player leaves and somebody else takes over their country later? If they take over before the end of turn there's no real difference so I'd say the game results count, at least for the other players.

What happens to the ratings of player(s) who join part way through a game?

How many turns can a country be uncontrolled for before it affects the game enough to invalidate the results for everybody?

Further, I think that if a country misses a Spring turn it makes less difference than if it misses an Autumn turn. What do others think?

The current system is awesome. It is the Ghost's baby, so he can do with it what he wants. Keep up the good work.

What happens to the ratings of player(s) who join part way through a game?

It is at present impossible to tell the difference between somebody who joins part way through and somebody who starts at the beginning.

Maybe with the data you are using, but the chat logs should show when someone takes over a position down to the season they took it over, Ghost. If you have access to them, you could write a SQL statement that joins and filters those games where you find the player took over the nation from their list. Of course, it would change the "zero sum" approach you basically have now because someone would get the GR points while the takeover didn't lose them.

I know that it would be possible from data on the site. I plan at some point-*beyond* my exams- to play with the data to see how much information that was once lost can be won back.

@TheGhostmaker:"It is at present impossible to ... at the beginning."
Oh. Ok. Thanks. I am surprised by that though.