@cypeg. D-points are not problematic. I like them the way they are.
D-points may be problematic only for who thinks they're an exact measure of someone's skill. Well, they're not, although skill is somewhat involved, a little.

So if the debate is about whether or not we can come up with stats that truly (or even better than currently) reflect a player's ranking across all variants and play types and (God forbid) SRG's, then I'm on the side that it's pretty much impossible (or still highly difficult in the case of better than current), and this discussion should cease (or be focused toward productive 'better than current' and doable steps that most of us would agree should be taken).

However if the question is can we add some highly useful, more informative, relatively easy to code stats, that don't necessarily indicate best overall player but rather best at say WWIV or 1v1 and that many of us would find more pleasing than being limited to the current stats, why would we not do so?

Also, cypeg indicated that D-points are problematic when used as stats, so you and he essentially agree on that point Guaroz.

@Leif, There is no current ranking system for skill. (My opinion is that D-points are not a reflection of skill). So, from my perspetive, we are trying to find an initial system for ranking skill.

I think the Ghost Rating system would be fairly easily to implement, could be adjusted to account for different map types, and automatically takes into account variations in skill.

@Ruffhaus, I doubt any significant number of players are using strawman accounts to pad their stats, but I have not really looked hard at this. As for winning against mediocre talent, there is more mediocre talent on this site than skilled. But, there is no need to blame the victors for this.

yeah, sorry. I should have differentiated better between my point of "lets forget the argument about getting a ranking system that completely accurately reflects player ranking" from the other two arguments of "lets get any kind of ranking system (that is better than D-points)" and "lets add useful stats that we'd all want to see".

I agree ghost rating would be easy enough to implement and tuning the equation would be where all the work would be (how or whether to balance the various parameters such as number of players, variant imbalances, likelihood of draw vs solo on a variant,1v1 issues, etc.) With even minimal thought in the tuning of the equation, GR would be way better than what we have already.

@Leif, Then I think we are generally in agreement. Getting a perfect system would be impossible, but it would be nice to have some system. Although, I don't think there would be any need to balance parameters. Balancing one thing will cause an imbalance somewhere else. I think it would open up endless controversy.

Devonian -

At the very least, any GR on this site would have to disqualify 2 player games and possibly 3 player games. Otherwise anyone whos good at tactics (and might suck at diplomacy), could fly to the top of GR by winning a lot of 2 player games. Is such a player better than someone who wins 1 of every 4 classic game (example)? Id argue not, but by not adjusting parameters, they likely would be ranked higher.

I may be an extremist, but I'd disqualify any game with 5 players or less. Not enough strategy, not enough diplomacy into them.
Btw, what is "GR" for?

Not necessarily. Using the formula I used earlier where the buy in is 5% of the player's current skill, a two player game between a noob and a skilled player with 1000 skill-points. The buy in for the noob would be 5 skill-points, and the skilled player would be 50. If the noob wins, they would net gain 50 skill-points, and the skilled player would lose 50. If they tie, the noob would net gain 22.5 skill-points, and the skilled player would lose 22.5. The skilled player would have to win to gain skill-points, and then, he would only get 5 skill-points. But, a tie or loss would cost him either 22.5 or 50 . That is a lot of work, and risk, to gain skill-points by playing 1v1 games against noobs.

On the other hand, if both players are very skilled, and both risk 50 skill-points, one player will go down 50 and the other up 50. If 7 skilled player all played against each other in a classic game, and one won, 1 player would get 350 skill-points, and the others would lose 50 each. To get the same skill-points from 2 player games, it would take 7 wins for each classic game won. Or in your example, a player who wins 1 in 4, classic would have to win 83.3% of all 1v1 games for the same skill-point increase. 83.3% is an incredibly high percentage. I have a pretty good 1v1 win percentage, and it is still only 74.5%.

Guaroz, Ghost Rating is a skill rating system that takes into account different skill levels when determining the relative skill. Winning against higher skill level players advances your skill rating much more than winning against unskilled players. Conversely, losing to unskilled players reduces your skill much more than losing to skilled players.

I think that looks like it should work. The only problem I can see is that games will get harder to fill as people avoid games where they are unlikely to make a profit, however this may also be a positive thing as people of roughly the same skill will generally end up playing each other.

Devo I like the Ghost Rating idea and I hope it's implementable.

But my point is that maps with a few players, resemble to Diplomacy but they're not. They're a different sport.
I'd love to play a ping-pong game with you, but I'd be baffled if you asked me to bet some diplomacy-skill-points on it.
So why should a noob have the chance to earn 50 skill-points in a game that has nothing to do with Diplomacy (aside the Map) ?

This site is not about diplomacy. It's about diplomacy variants. I think all variants should be included in the rating. If someone wishes, they can go to webdip and get their skill rated there. Web-dip only rates the classic diplomacy games.

Besides, I think a 1v1 game tests a players skill more than a game of chaos, or any gunboat game, or most games with unusual rules. If we start trimming out variants, then we need a rating for each variant. I realize 1v1 games only test "tactical" and "strategic" skills, but these skills are essential in the game. The skill of "diplomacy" is only 1 of the 3 skills needed to be successful.

Also, you don't really "bet" skill-points in the way I think you are suggesting. Your skill dictates how much is at stake in each game. You do not have a choice to "bet" more or less. It is always a fixed percent of your current skill-points. You can ignore the rating if you want, and let your skill-points land where they will. You can never go to zero, and there is not a minimum buy in.

I must respectfulky disagree with your assessment Devonian. Whike diplomacy is one of 3 main skills, it is also by far and away the most important skill. The best tactical and strategic player will almost akways lose if his diplomacyy sucks. On the contrary, a good diplomat can make up for shortcomings eksewhere with good diplomacy. No one player can win on their own, thats the whole idea behind the game. Therefore, any game whoch allows one person to win on their own is not Diplomacy in any way shape or form. It would be like calling basebalk and cricket the same thing. Similar but defonitely different.

Regarding other special ruke games, at least those keep the same general layout of requiring diplomacy to win. 1v1 does not. Arguabky neither does 1v1v1 if youre lucky. To me, any game which takes diplomacy out is no longer a variant. It is a different game using similar rules. After all, how can it be a DIPLOMACY variant with no diplomacy?

Oh and sorry for bad formatting and typos...damn phone...

@Devo.
1) what Drano said. It's a different sport, I like it very much, it may be very difficult (you know I know, lol), but it's not a Variant.
2) No strategy in 1v1, sorry. Only tactics. There's no strategy because you can't choose whom to attack and you don't have to guess who will attack you. Both cases, it's the other. There's noone else! The only strategy in a 1v1 is to defend from and to attack the other.
And 1v1v1 it's not much different. As I said, IMO in games with 5 players or less there's not enough strategy, not enough diplomacy. You may not agree that the strategy is not enough in an Ancient Med, but I don't see how you can say that there's some strategy in 1v1.
3) Yeah ok. Whatever is the bet-system (choosen by players or automatized) I don't like someone gaining diplomacy-skill-points playing a game of a sport that have just tactics in common with a Diplomacy game. No strategy, no diplomacy.

@ drano,
I am open to respectful disagreement... but not bad formatting & typos. :)

I agree that the skill of diplomacy is the most important skill. But, I won't go so far as to say it is the only thing that matters.

I believe that the art of diplomacy is the ability to convince an opponent to take an inferior strategic or tactical position. So, in this sense, there can be no "meaningful" diplomacy unless there is tactical or strategic aspect to the game.

For example: The Pure variant has no tactics and no strategy. These games often are a continuous series of dislodgements and retreats until someone cannot retreat. But, no-one will intentionally agree to be disbanded without a retreat. So, the game reverts to trickery, deceit, and out-guessing an opponent. This is not diplomacy, even though it is called "pure" diplomacy. In my opinion, a winner cannot seriously claim to have better diplomatic skill.

@Guaroz,

Yes there is strategy in 1v1.

For example, the decision to I mount an attack by land or by sea is a strategic decision. Or, when the enemy is trying to make an end run, should I try to stop it, or should I take advantage of his weakness in the middle. The actual moves and supports are tactics. The overall plan is strategy.

Here is another strategic decision in 1v1:
A fleet is dislodged in Western Med. Should it retreat to Tyrr and defend, or go to Mid Atlantic Ocean, and go on the offensive. A lone fleet in MAO has little tactical advantage, but huge strategic value. Whereas, a supportable fleet in Tyrr has huge tactical advantage, but little strategic value.

Though you are right in tactical advantage, I completely disagree with the setting of 1v1. For two reasons, a) map is not balanced (with the exception of DUO) and b) I see 1v1 as a raise to 18 centers and not really much of tactics. IF however, 1v1 was elimination type then I may reconsider.

I recently played a 3 man game: "843 the treaty of verdun". which I found very satisfying both in diplomacy and tactics. thus for me any map challenging enough from 3 players up should be included in any new stats.

*race not raise

@Devo. You are Stalin. Here's your game against Hirohito.

**Strategy.
1) 1941-on. Wait and see what FD does.
2) August 6, 1945 "Oh shit, it's time to attack!"

**Tactics.
August 8, 1945: Send some guys to Manchuria and South Sakhalin.

**Diplomacy.
1) August 8, 1945: Declare war to Japan.
2) August 31, 1945: "Hallo!...Harry?... Hi...yeah I know... listen...do me a favour... Many Russian boys died because of Japan, we did our share. So, day after tomorrow, could you please make Shigemitsu sign that South Sakhalin gets back to us? A small reward for our big efforts? Mm? Uh, ok, many thanks and.... What? Missouri is in the Tokyo Bay? Fantastic, I'll send Derevyanko with a camera! Very welll... Cheeeers♫... give my regards to Bess!"
:click:

- - -

Now, if you ask to a Russian General who was sent in Manchuria, he'll tell you that also his job can be divided in tactics and strategy. But his strategy is not the same of Stalin's. The General couldn't decide when to make war, and he can't decide when the war with Japan is over. It's not a real strategy, it's rather high-tactics.

In a 1v1 you play as a General who was ordered to attack 1 enemy. You can't decide to delay your attack to that enemy. You can't decide to stop that war while your enemy is committed with another enemy. You will never think that that war is strategically wrong because it benefits someone else more than it benefits you. And you can't try to change your strategy attacking someone else. The only strategy in a 1v1 is to defend from and to attack the other General. Always. Forever. You can't change it.
So...It's only tactics. With some high-tactics, if you like. :)

Maybe we should play a 1v1 game. You can use your tactics. I'll use my tactics and strategy. :)

ahahah :D yes... ok! But you can use any strategy you like except attacking me! ;-)

I wouldn't mind the idea of the Global Ranking system, but the GR is designed for classic map-only, and... no matter how you mean to adapt it here, there will be always critics about how to take into account:
VariantID #1 instead VariantID #2 or VariantID #50
gunboats instead of full press games
2 players games instead of 3, 7, 34 players games (the most discussed issue)
games with NMRs/CDs intead of flawless games
SRG instead of standard games
games with standard win condition instead of games with altered win conditions
password-protected games instead of public games
.... it's a mess.

Moreover, the GR system (correct me if I am wrong) works on a separate website.
What % of users here bother about user stats? How much will that % drop after the user stats are not integrated into this website anymore?

Anyway, despite those imperfections, basically I am in favour of the GR idea.
In the instance that a GR system will be considered not worth to be set up, I will stand behind my initial thought:
use the current stat system, just separating fields coming from 1vs1 and from multiplayer games in two different stats on the user page.
I am aware, it's an incomplete system, but it's simple and more familiar to everyone.

One option would be nice to add to the search page if you want to see beyond the current stats without scrolling down tens of games of the user page: searching games for userIDs.

instead of the current option
-All
-Joined games
-Non-joined games
changing with
-ALL
-Joined by userID=XXXX
-Non-joined by userID=XXXX
(either userID=XXXX is yours or someone else's ID)

I am no math expert, but can we do a formula where it takes into account all the centers of a map, solo requirements, how many centers obtained by player, and number of players.
So a classic has 34 centers, solo is 18, I control 10, 7 players.

Decima, Yes, there will always be critics. But does that mean the idea of a global rating system should be abandoned? I also would say that a global rating system should truly be global, and include all game styles. Otherwise it is not a global rating system.

But, aside from a rating system, better game searching would also be nice. (The search by user id is already available. You just have to go to the player's profile, and search their games)

But, I would like to be able to search by groups of games also. For example, did a search, and see you have played 2 GvI games, but if I want to find out how many 1v1 games you played, I have to do 8 searches on all 1v1 variants.

since the games go to our profile automatically then it shouldnt be difficult to insert them in an excel type of format

LOL @me Devonian, I think I never realized that "search" on the user page was actually a search-engine link... ok, forget my last sentences.
Anyway, I'm fine with the GR idea; I just pointed out those that look to me as the main problems