*not Top 10, but close.

One factor thats not taken into enough account in the present system of awarding VDip points is that the points are shared among all surviving players in all non-solo games, giving no weight to the number of centers captured. In some ways this makes sense because if a nation can survive, then they have indeed earned a draw, and should get some regognition of this. And no player should be encouraged by the points scoring system to be out there dumping solos. But it's more complicated than that, and it creates an atmosphere where we play game to go out and grind out the small powers at the end of games, mercilessly shooting the wounded. This is also a big reason why so many players quit their games once stabbed. If you have 15 centers and another player has 2 centers, what is your motivation to allow this player to survive the game? Why should that players score equal points?

The solution to this situation is to award a base score for nation perfomance in terms of solo, draw (taking into account placing in the draw - 1st, 2nd, 3rd, 4th, etc), surviaval, elimination with a bouns of some sort awarded for each supply center owned. Now because some variants have over 200 supply centers, I'm not suggesting that a supply center owner in a WW4 game is equivalent to one inered on the standard map, but there should be some regognition of Maybe the supply center bonus is calculated as a percentage of the total centers available and then also wieghted agaisnt the number of players in the game.

I'm not suggesting the following scoring system as a direct conversion for use here, but one of the popular Diplomacy scoring systems used in tournament play including the once upon a time Worldmasters competitions was a very effective method for encouraging competitive play and survival with high risk and reward for solos, while taking into account the value of high SC accumulation and observing the merits of meager survival positions. It's my opinion that this methodolgy of scoring give players somethign to play for even once a game goes bad, and that by creatign something similar we would avoid a lot of NMRs and CDs. A sample tournament scoring system:

Solo Victory = 100
1st Place (draw) = 32
2nd Place (draw) = 24
3rd Place (draw) = 16
4th Place (draw) = 12
5th Place (draw) = 8
5th Place (draw) = 4
6th Place (draw) = 2
7th Place (draw) = 1 point
Elimination = 0
Suffering a solo = 0

Placeing points are combined and split evenly in the case of a tie. And added to this is a one point per supply center center bonus held at the end of the game. So a 17 SC board top/non solo is worth 49 , (32+17). 17-17 draw is 37 for each player (32+24)/2 = 28+17 for 45 each. And at the bottom of the scale a 4th place 2 SC nation that survives the game taking place in a draw gets 14 (12+2).

The value to this is that it gives players a legitimate reason to play o, and play competitively even when they cannot win. It also give everone there encouragement to both go for a solo while at the same time doing everything they possibly can to prevent a solo form occuring on their watch. And perhaps avodis incidents like we saw in 'Whatever Happened to WW3' where one player gave his friend a solo and still scored significant points. A solo in Diplomacy means one win, and 6 (or 34 in the WW4 case) losses. There is no such thing as second place in a solo situation, because a solo assumes that the player with 51% of the supply centers will go on to conquer the rest of the board, and the convention at this point is akin to chess where you accept simply accept the mathematics of eventual defeat and lay down your king.

Decima,

"Using the current formulation, by a player’s point of view:
After a duel my rating can change +/- 15
After a classic my rating can change +/- 15*6 = +/- 90
After a KW_901 my rating can change +/- 15*14 = +/- 210
After a WWIV my rating can change +/- 15*34 = +/- 510"

I was thinking exactly the same thing. By treating every game as a 1v1 matchup, there is way to much emphasis on both large map games, and winning. For example:
if 35 brand new players all played a WWII game, and they never played diplomacy before, and one player won, that player would gain exactly 255 . That would list that person as the 18 best player!

Another problem is that it does not reward skill in the same way that the games were originally played. All games are effectively WTA. There is no PPSC.

Ruffhaus,
your idea is users-skill-independent: you propose to appoint 100 to the winner no matter who were the users he defeated.
We’re seeking for an users-skill-dependent formulation… else, we have a formulation fairly similar to the current D-POINT one.

Devonian,
indeed, the Won/Draw/Lost rating algorithm models a multiplayer game into a series of WTA isolated duels, regardless the way the game was played, PPSC or WTA.

RUFFHAUS_8 wrote :

"And perhaps avodis incidents like we saw in 'Whatever Happened to WW3' where one player gave his friend a solo and still scored significant points."

I am the player you are talking about. But you are distorting the facts, RUFFHAUS_8.

1° In this WWIV game, I was playing Egypt. I had a perfectly faithful alliance with the player playing Nigeria. That player was a TOTAL STRANGER to me : I had never communicted nor played with him before that game. As a consequence, THAT PLAYER WAS NOT A FRIEND.

2° At some point, I was leading in the game (yes, with the supposedly weak Egypt !) and Nigeria was holding the second place. But none of us couldn't make a solo (which required 50 SCs in that game) without betraying the other one.

3° Then, in order to avoid that any other player wins a solo victory, I proposed to my ally that I would give him some SCs so that he could THREATEN to win a solo victory. We wanted to force the other players to vote "Draw"

4° BUT, some players, YOU BEING THE FIRST OF THEM, REFUSED TO VOTE "DRAW" because a "Survive" would reward you more points than a Draw !

Indeed, on October 26 2012, you wrote in the Global Press of that game :

"The bottom line is that African demand doesn't offer the other players any incentive to accept for it. We all accept that a 50 SC solo isn't a real solo, so there's no shame in that outcome. I'll take 64+ points over 32. Frankly I'll take 64+ points of the 8 player draw which gives me 48 . You're doing me a favor. So thank you!"

5° Finally, in order to avoid that any other player wins a solo, I allowed my ally to capture more of my SCs so that HE could win instead of another player.


So, RUFFHAUS_8, you have MOST DISHONESTLY distorted the facts when you wrote :
" And perhaps avodis incidents like we saw in 'Whatever Happened to WW3' where one player gave his friend a solo and still scored significant points."

Also, in your previous message, you wrote :
" It also give everone there encouragement to both go for a solo while at the same time doing everything they possibly can to prevent a solo form occuring on their watch."

Why are you writing that if you don't yourself do everything you possibly can to prevent a solo form occuring ? I can see only one answer : YOU PLAY FOR POINTS ! Like most players do, actually.

That is why, I have written in my first answer in this thread :
" 3° By the way, has anyone thought that this new rating could possibly - and most probably will ! - pervert the play of many players ? For example, I have seen here a game (in the only 4 that I have finished here !) where several players refused to vote "Draw" and prefered instead to be defeated because a Defeat would reward them more DPoints than a Draw ! That was obviously not the spirit of a Diplomacy game, at least in my opinion !"

As a conclusion, I would finally like to say that I think that any new rating system is in fact a very bad idea because :
- I do think that it will be impossible to find one method of calculation that is totally fair.
- Much more important : it will pervert the play of many players !
- As it has already been written here several times : another problem is that it does not reward skill in the same way that the games were originally played. And that is MOST unfair !

Finally, if there is a new rating system, I would please like to suggest that the points are called :

EGO POINTS

But at the moment if 35 brand new players join a 100bet WWVI game the winner can gain up to 3500 Dpoints hitting the top score. And once this place is reached there is no easy way to loose the points. If a player scores 510 in a WWIV it will dump relative quick if he loose (or even draw) some more games...
Maybe we should require a minimum of 5 or 10 (non 2-player) games played to get a visible rating.
After the first few games his rating will adjust quite fast.

The problem is oli, nobody is perfect, the best player doesn't always win. If you happen to do well in a WW4 game with Talrus, Orpheus and Devonian in it your rating will sky rocket. At the moment ratings will jump or fall too quickly, which was a problem in the early Ghost Ratings.

The solution was any change in ranking was divided by 17.5 so consistency matters more then a few wins in big games.

On that note I am surprised Drano069, who has a win/draw record of 90%, isn't higher on the list.

Also looking at the current algorithm I think I found a mistake (please correct me if I am wrong):

Suppose both players had an WDL of 1000:
So E1 and E2= 1000

Re1 = 1 / ( 1 + ( 10^( E2 - E1 ) / 400 ) )
Re1 = 1 / ( 1 + ( 10^( 1000 - 1000 ) / 400 ) )
Re1 = 1 / ( 1 + ( 10^(0) / 400 ) )
Re1 = 1 / (1 + ( 0 / 400 ) )
Re1 = 1 / 1
Re1 = 1
Re2 = 1

Shouldn't the Re1 and Re2 both be 0.5 not 1, since the expected result of 2 people with equal WDL scores be a draw?

If continuing with this example say they get a draw:

Edif = 15 * ( 0.5 - 1 )
Edif = 15 * ( -0.5 )
Edif = -7.5

E1 = 1000 - 7.5
E2 = 1000 - 7.5

Unless I have made a mistake (please tell me if I have), 2 players of equal WDL, if they draw, will both decrease in score. I thought if WDL is tied and the game is a draw, both players should stay the same.

Am I making a mistake somewhere?

Oli,

"But at the moment if 35 brand new players join a 100bet WWVI game the winner can gain up to 3500 Dpoints hitting the top score. And once this place is reached there is no easy way to loose the points."

That's exactly why we want to get away from using the current d-point system. This system does not fix that problem.


"If a player scores 510 in a WWIV it will dump relative quick if he loose (or even draw) some more games..."

The maximum a person can lose in any game is 15 . So it would take 34 consecutive losses to get back to neutral.

The system I proposed WOULD allow an higher score to drop down quicker, but this system does not.

@fasces 10^0 = 1

My bad:
so to fix that mistake:
Re1 = 1 / (1 + ( 1 / 400 ) )
Re1 = 1 / 1.0025
Re1 = .997506...
Re2 = .997506...

Edif = 15 * ( 0.5 - .997506 )
Edif = 15 * ( -0.497506 )
Edif = -7.46259351

This time I used a calculated rather then by myself, but based on how I interpreted the formula, your WDL would still go down despite this being a draw with someone of equal rank.

*calculator

Retillion,

"That is why, I have written in my first answer in this thread :
" By the way, has anyone thought that this new rating could possibly - and most probably will ! - pervert the play of many players ?"

I agree. That is why whatever skill-point system we create, it should apportion skill-points the same way the current D-point system works. The problem with the current d-point system is not the apportionment of d-points. At least, any perversion, has already been adapted by the community. The problem is that anyone can pick the d-point value of each game.

Two possible solutions:

1 - Each game is always worth 5 skill-points, and apportion them according to the current d-point system. This does not adjust for playing against skilled players, but is better than the current d-point system. (OK but not great)

2 - Each game is always worth 5% of each individual's skill-points. Then apportion them according to the current d-point system. This adjusts for individual for playing against skilled players. A skilled player gains less by beating a noob, but a noob gains more by beating a skilled player. (very comparable to the Ghost Rating system)

@fasces: / is bevore ^.
So it's:
Re1 = 1 / ( 1 + ( 10^( 1000 - 1000 ) / 400 ) )
Re1 = 1 / ( 1 + ( 10^(0 / 400)
Re1 = 1 / (1 + 10^0)
Re1 = 1 / 1 + 1
Re1 = 0.5

Fasces & Oli,

I think if "Edif =15*(Rr - Re)" is supposed to be Edif = 15*(Rr-Re1), then I think the formula's are right.


But, I think the rationale for using these formula's are wrong.

1. It does not use the same measurement as how the games were played. So, the historical data gives and takes skill where it shouldn't.

2. It places too much weight on large games.

3. It places too much weight on winning.

4. It allows for rapid jumps in skill rating, but only gradual declines.

5. By using a different measurement, play will be perverted to conform to the new scoring system.

@Devonian:

1. ??

2. Large games still tend to last much longer, so I'm all in favour to wigth them much more.
We tried to nerv the smaller games with a maximum of DPoints to bet, but if you set a new limit there is a new way around this. Matchin 1on1 makes a 13-person game weigth the same aprox as 2 6-7 player games. And this looks ok to me considering the time commitment.

3. I don't think so. A 3-way draw in a 7-player game scores 2-draws and 4 wins for each player in the draw. A win in a 7-player game with 2 survivours gives the winner 7-wins, the sorvivours each 1-loss, 1-draw and 4-wins. Still not shabby.

4. As fast as someone can reach a goot spot in the lists, as fast he can go down if he does not perform that good. To keep at the top you must _constantly_ beat the players better than you. Drawing to players with a lower score will put your score down. This basically forces the player better to play the game as intended (play for the win).

5. We can say the same for the DPoints, as people see a better result in a survive than in a draw at the moment.

And if you look at the calculation of the ghost-rating (esp. the formula at the bottom), it's much more similar to the Elo-rating, than to a fixed part of your skill-points used as bet.
https://sites.google.com/site/phpdiplomacytournaments/theghost-ratingslist

I think with the knowledge I gained through this exercise I should be able to implement the GR here too (Had no clue how this formula works at the beginning of this exercise)....

That said, I still like your approach too, and I will give it a try too.

That makes more sense, the brackets in the first line are confusing it should be:
1 / ( 1 + ( 10^( ( 1000 - 1000 ) / 400 ) ) )
Your formula was just missing a bracket, which caused both me and my calculator to get a result that was higher then what we were suppose to get.

My New problem is the size of the map is a bigger importance then who you face.

Lets play 2 examples. A 1v1 with Talrus (person with highest score) facing an unnamed rookie vs 3-player game between 3 unnamed Rookies. In both examples an unnamed rookie will solo. Obviously beating Talrus would be a challenge.

The ELO ratings calculator predicts that Talrus should be able to win 89.10% of the games against someone with an WDL of 1000.

Defeating him would boost your WDL by 13.365.

Now lets look at the 3-player map. In theory the probability of losing is 2 in 3, or 66.67% of the time. Based on this it can be concluded that playing nothing but rookies in a rat wars map is easier then playing Talrus 1 on 1, right?

From how I understand the calculation, it would be:
In all cases (Rr - Re)*15=Ediff

Rr = 1, Re = 0.5. Therefore Ediff=7.5

If I understand it correctly because you are defeating 2 simultaneous opponents. Your RR is boosted by 15.

I would say beating the best player on the site 1 on 1 is harder then beating 2 rookies in a 3-player map. In fact the odds of beating Talrus is 1 in 9, but the odds of beating the 2 rookies is 1 in 3.

Thoughts?

Oli -

Just my opinion, but shoukdnt survive = defeat in a game with a solo? Why should someone get 30 wins, 3 draws and 1 loss (in a hypothetical 35 player game with a solo and 4 survivors) when they lost just as much as a player who was defeated did. It seems really odd that a 1 sc survive would yield 30 wins, whereas a defeated player gets 30 draws instead. What that does is encourage small powers who might be killed to throw the game for a HUGE ranking bonus.

Does this list update automatically as the games finish? Because gameID=10863 just ended and my Velo didn't change...

My solution would be to further simplify the formula and see what this basic formula does to the result:
Re = E1/ (E1 + E2 + E3...)
Basically if its a 3 players map, then it would go up to E7.

Then we can simplify the middle:
If S1 = won, Rr = 1
If S1 = Draw, Rr = 1/n, such that n is the number of players in the game.
If S1 = survive or defeat, Rr = 0, since his is a WTA model, survives=losses.


Ediff would still equal 15*(Rr-Re)

And the New ranking would still be old ranking + Ediff.

Thoughts?

I can't do any example scores on this system because I don't know how high the best player on the site would score on this one. I think it would create a larger range of scores.

I think the problem with mine is that scores wouldn't change fast enough. In order to be considered twice as good as an opponent you would have to beat them 66 times, in otherwords my system places an importance on # of games played.

I would say rather then Ediff=Rr-Re*15, I would change it to something like Ediff=(Rr-Re)*1000/x, where x is the number of games played. Problem with this is winning 1 game and never playing against would put you at the top of this list.

So maybe *1000/(x^1/4) or something along those lines. Thoughts?

Drano also points out a huge problem with the current ratings

Oli,

1. For example, if a player is trying to maximize their d-points, they might take a 14 point survive instead of a 3 way draw. The d-point system rewarded the survive over the draw. But, this new system rewards the draw over the survive.

The person was playing for more points AND SUCCEEDED. How can we now say the person failed because we changed the way we measure success.

5. Yes, but PPSC has been part of both vdip and webdip from its inception. It has become an accepted part of the game by the vast majority of players.

Retillion, It's too bad that you find me unpleasant, but that is your choice. It's likewise unfortunate that you find recollection of the facts of your past behavior to be unpleasant, but you own that, and none of us can change that. Perhaps my point had little to do with the discussion, but you're the one that mentioned it. Why did you bring up the Clockwork Angels game, and more specifically, why tells us all how you wiped out Cypeg? Was there some value to the discussion at hand? Was there some value to the level of mean spirited specificity there? No.This was just you crowing about your accomplishments in a game you formed with pre-chosen allies and enemies, and one that you also prevented certain players from joining. Given those facts I'd hardly call it a stellar achievement on your resume. But you can do so if you wish to. It's a bit far off topic though.

I am not blocked by several players. You are the only active player here that I am blocked by, and given your dubious ethical boundries, I'm rather thankful that that your too afraid to go up against me.

Decima, the scoring exmaple that I posted was only intended to be considered as part of the overal ranking system. The specific points numbers used there are only there to illustrate the example. I am absolutely in favor of including some factors that recognize strength of opposition. I posted the system mainly to demonstate the value of recognizing the placement of nations in a draw and the total supply center ownership at the end of draws. Maybe the system I used puts more emphasis on this than it should, and once again the specific amount is not relevant. The salient point is that such a system offers incentive to weakened nations to fight on, while also allowing the mores successful nations alternatives to simply grinding up all the little guys before calling a draw.

I'm in game now with a player that is down to one or two centers and pleading for mercy. I'm most likely going to eliminate this player not out of mean spirited maliciousness, but out of simple practicality, and the current scoring system. I see no reason to allow that player to share completely equal points with me by allowing him to survive the game and be included in the draw. I've played and GM'd this game more than a few hundred times under the scoring system that I proposed, and I can tell you assuredly that this brings a richer and more enjoyable game with fewer NMRs and abandonments, and much more exciting and entertaining games.

I think people woudl be surprised how much you can learn about this hobby by playing out mercy positions, and if we adopt a scoring system that offers some possible potential to earn points from them, that all of the games we enjoy here will be the tbetter for it.

" I have seen here a game (in the only 4 that I have finished here !) where several players refused to vote "Draw" and prefered instead to be defeated because a Defeat would reward them more DPoints than a Draw ! That was obviously not the spirit of a Diplomacy game, at least in my opinion"

&

"Just my opinion, but shoukdnt survive = defeat in a game with a solo? Why should someone get 30 wins, 3 draws and 1 loss (in a hypothetical 35 player game with a solo and 4 survivors) when they lost just as much as a player who was defeated did. It seems really odd that a 1 sc survive would yield 30 wins, whereas a defeated player gets 30 draws instead. What that does is encourage small powers who might be killed to throw the game for a HUGE ranking bonus."

I completely agree. Survives should, if at all, be only minimally awarded points. No player should desire a survive (which is really a defeat) over a draw (which is better than a defeat, but worse than a win).

Taking into consideration the BS Retillion is saying brings up the problem of reward over more highly skilled players.

But first I think that
1) Solo either in PPSC or WTA should be the highest score
2)- a draw in a PPSC game should be less rewarding that a draw in a WTA ?
3) survive = defeat when someone achieves a solo.

and to answer Retilliion's "For your information, in that Clockwork Angels game, I personally fought and eliminated (with my allies of course) Cypeg."

a) Personally fought??? looooooooool you barely hit me once in that game.
---) my defeat was the result of a combined Song - Japan - Indonesia-Australia attack which only managed to stall me. They succeeded only whenl a 5th person, India (Gary123) decided to stab me out of the blue. SO if 5 men need to take Thailand down then what can I say! well done for "personally" fighting me and getting all that credit points in the new system!

But I like how you brag about a Draw in which first you took advantage noob players around you and then just bullied the rest.

In that game I picked Thai to play test it. If I had taken Germany you will not be talking now.

@RUFFHAUS 8:
How would you change the 100-points based on the number of players.
In 2-player games someone could easy score 1000 and more in a matter of hours.

As someone mentioned above, due to the point system, several players play for the draw which maximises points. However, the objective of the game is to solo. This is the only accepted rule in standard. Thus a solo should be awarded heavily.

Perhaps instead of making a formula about skilled players or not why dont you categorise them according to their wins-draws i.e. those who have 50% win iinto one categoty those who have 40% win+draw get to be in that category of skill