Which is why we have different scoring systems to choose from. And by abolishing scoring and ratings, you do not take away the fact that different people have different opinions on how the game should be played. In fact, I would argue that it would make matters worse. At least if you use a scoring system, you have an objective measure of how the game should be played: just maximize your point gain. Without a scoring system, this objective measure does not exist. Imagine we were to abolish all scoring systems and former 'PPSC advocates' would play games with former 'WTA advocates'. That would lead to frustrations and to people accusing each other of not playing the game how they should.

Now if you have already (a) scoring system(s) in place (WTA and PPSC), I think it makes sense to implement a rating system, too, but such a rating system should be in line with what the players want. When it comes to ratings, I do think that there is more consensus in the community. I think everyone can agree that headhunting is bad and that the Divided States solo example I outlined is ridiculous. There are of course people who complain about the existence of ratings altogether, but honestly I do think, though I can't look in their minds, that their complaints would almost fully go away if the rating system was improved. This is the only place where I have seen people complaining about the mere existence of ratings. On webDiplomacy, no one is doing that, and I think the underlying reason is that the implementation of their rating system just has had far fewer unintended consequences, and generally works better for their site.

I've said this before, but I'll say it again. I think the different ways people play the game is a huge part of the fun of Diplomacy. There is no particular way to win, some people find winning solo to be the most rewarding, and others enjoy drawing with close allies, I have done both myself.

I won't talk about what the rating system should be, or whether it should exist. But I think it isn't a really bad problem. Rating system or no people are going to play the game a certain way. Like the "draw" versus "solo" argument that exists within WTA games alone without even bringing PPSC or another scoring system into consideration. That debate would exist no matter what scoring system was implemented. Diplomacy is a game of relationships not points. There will always be people who prefer a particularly good stab and others who would rather stick it out with great allies.

I think this trait is something pretty unique to Dip, and something we should embrace as a positive attribute of the game we all love.

@Uber -

Except that's the problem. There's a group of people (and I'd guess a fairly sizable one) who don't think that having people who just "stick it out with great allies" is a positive attribute.

To be clear, I'm differentiating people who are risk-averse (and therefore less willing to stab for solos and risk someone else soloing as a result) from those who play with game-long allies with the intent to draw.

It's the second group people have (more) of an issue with. It's not uncommon to see groups of 3 play game-long for a draw on classic Dip here on vDip. How is that fun? You have 42.8% of the players on one team, plus the element of surprise since it's not exactly known by everyone that they're all a team. For reference, that'd be like 21 players teaming up for a draw on Divided States.

The issue people have with it is that it's changing the very fiber of the game. Turning a game where the goal is to get 18 SCs and solo into a co-op campaign game. The two views simply are NOT compatible, and the people who suffer when they mix are the people who are playing to try to win a solo, since inevitably, 1 loses to 3.

@drano,

I see your point. But I still think that the ability to accommodate different play styles is probably one of the best things about Dip.

But wouldn't you say that many of the players that hate that play style were victims of it? Maybe the "drawers" would see your point of view if they were victims of the same style.

I to look down on the practice of planning to draw with people from the beginning of the game, but my point is that a different rating system won't change that. Players will always want to draw with other players who they really enjoyed playing with even if that means forsaking an obvious solo route, you can't change that by getting rid of PPSC.

Finally I'll say it again: Diplomacy, much like life, is a game of relationships not points, it is not "about" getting to 18 SC's, it's about the friends, and enemies, you make along the way. This game is much more fun when you have great friends and enemies, players who are great allies but will mercilessly stab you in the back when it favors them and will be a gentleman when you do the same to them.

I would much rather play a game with good friends and enemies that ends in a sucker draw then a game with jerks and whiners that ends in a solo by myself. Maybe we differ in that respect, but I think I have a good point.

@Uber -

"I would much rather play a game with good friends and enemies that ends in a sucker draw then a game with jerks and whiners that ends in a solo by myself. Maybe we differ in that respect, but I think I have a good point."

Believe it or not, I'd agree with you. Jerks and whiners suck, and playing with them sucks. But it's not a black and white issue. Why can't we play a game with good friends and enemies that still involves mercilessly stabbing an ally in an attempt to solo, and then desperately making up with that same person when someone else threatens a solo?

My whole viewpoint is actually pretty much summarized by you in your second to last paragraph.

"This game is much more fun when you have great friends and enemies, players who are great allies but will mercilessly stab you in the back when it favors them and will be a gentleman when you do the same to them."

^^^that is good Diplomacy. It's when people refuse to ever stab an ally, EVEN when it would obviously advantage them and/or lead to a solo for them that I have issues. And we seem to see that a decent amount around here. It's not necessarily a ratings issue, but more of mentality issue I guess.

"Why can't we play a game with good friends and enemies that still involves mercilessly stabbing an ally in an attempt to solo, and then desperately making up with that same person when someone else threatens a solo?"

You are free to do that and I would love to join in. Just don't complain when someone won't do that because they played a great game with their ally, in my opinion that' whiny.

"It's not necessarily a ratings issue, but more of mentality issue I guess."

That is exactly the point I am making. If you want to change players minds about how to play the game complaining about the rating system is not the way to do it.

Mercy, I think your proposed reform is a pretty good way to improve the current system. Interesting that a thread on 100 SCs has become about everything else.

We've gotten on a tangent. I still firmly believe there's an issue with headhunting and the rating system. I just happen to agree that THIS issue is not caused by the rating system.

"You are free to do that and I would love to join in. Just don't complain when someone won't do that because they played a great game with their ally, in my opinion that' whiny. "

Would it also be whiny if a soccer player complained about his teammates who only cared about passing the ball to each other as many times as possible instead of trying to score goals and win the game? Of course not! Because those two people aren't playing soccer! They're playing a game that resembles soccer but that is completely different because they've changed the objective.

It's the same here. When people change the objective of Diplomacy from soloing to drawing with their allies it's a different game. And most importantly, the other players DO NOT KNOW they've changed the game. It's unfair to people who are trying to play the game to solo per the objectives on the rules to suddenly have 2 or 3 or even 4 people playing a different game against them! If people want to change the game, it needs to be known up front.

Like I said before, it's understandable that at times, people view a solo as too risky and decide to draw instead of risking someone else soloing. It's when people play from the start with zero intention of ever trying to solo that it's an issue. How fun is it to play a classic Dip game where France, England, and Italy all ally immediately and plan to 3 way draw? Because by the time Germany figures that out, he's dead. And how is it fair to Germany that he literally never had a chance to ever influence them since they're dead set on the 3 way draw?

I agree with you at least partially. I think that the losers of a game will always be upset no matter how they lost, so people who complain about such things were probably victims of it themselves.

But you're right, this has nothing to do with the rating system and no rating system will change it. Try to change peoples understanding of the game instead.

The point I was trying to make is that Diplomacy is not a game of trying to score goals or trying to pass the ball as much as you can; it is a game of having fun with the people you are doing it with.

"The point I was trying to make is that Diplomacy is not a game of trying to score goals or trying to pass the ball as much as you can; it is a game of having fun with the people you are doing it with."

I have to disagree completely with that. That can literally describe every single activity or game. The object of playing something is to have fun yes, but there must be a set of rules/objectives for people to understand how to play such a game. It is a VERY slippery slope to claim that just because you are playing g to have fun, you can play however you want to as long as you're having fun, and people who disagree are just whining.

To go back to soccer, people play to have fun right? But they also understand the objective of soccer is to score goals, and ultimately, try to win the game. To use your idea of playing to have fun with the people you are doing it with, I could do whatever I want on the pitch, and claim it's "right" since I'm playing for fun.

And in some cases that's true!! Very little kid leagues, or rec leagues might not care! But it has to be known up front that one isn't playing by the known objectives of soccer.

Same goes for Dip. If people want to claim that permanent game-long alliances resulting in large draws are their objective, fine! You just gotta tell people beforehand that that's the way it's being played, since that is NOT the basic objective of Dip, which is to gather 18 scs and be declared the winner. Playing that way and not telling others is deceptive and completely not right.

I think you may have changed my mind.

Let's take another example; chess. In chess the "objective" objective is to checkmate your opponent and there is a set of parameters to get there(e.g. knight moves like this, queen moves like this). The "subjective" objective is to have fun(which is shared by every game out there, other than educational "games", but that's another topic for another day). I used to play chess with my father when I was a child and I would ask to play "off the board" which meant we would push the pieces around and knock each others men over. Loads of fun, but it was not strictly chess.

In the same way a solo at 18 SC's is the "objective" measure of "winning" in Dip. The "subjective" objective is the same of course. Some players play "off the board" and go for draws with great allies or go for as many SC's as they can get in a PPSC game or try to color all the neutral non-SC territories their color.... et cetera. This of course can be fun if all players know what's going on, but if I were to go to a amateur chess competition and insist on playing "off the board" I would certainly be kicked out.

This leads me to conclude that, while playing "off the board" is fun, the game needs rules.

I still think that Diplomacy is a great game of relationships and cooperation(or lack there of), but it still needs its basic rules to function in order for the game to work properly. Demanding that a draw is an acceptable "win" condition is tantamount to deciding that fleets can move inland and armies can fly over Switzerland and Russia gets another SC in Livonia without previously deciding that it will be like that with every player in the game.

Thanks for this productive discussion drano. I really appreciate having my opinion changed like that, it reminds me to be humble about my other opinions.

As some background history on the current ratings system for those who were not here, or cannot recall, there was an initial precursor algorithm, which only tracked success of players. A wave of people rushed in complaining that the system did not penalize elimination, nor did it address the relative strength of opposition. Instead of actually having an analytical discussion about the topic, the ratings system was immediately changed to the dogshit system we have now. Yes, some of you are going to wince, and say, why are you calling it "dogshit", it's not that bad.... It's worse. At the time VDiplomacy was thriving, and had dozens of high quality games to join at even given time. And even when you wanted to try a rare, or large variant, a game could be set up realities quickly. I'm sure that there are other factors at play, but the ratings system is responsible directly or indirectly for the departure/hiatus of a large number of core players. The system was railroaded in by Guaroz (ironically now also departed), who had Oli's ear and confidence, even when he was terribly wrong, like this time. And no amount of petitioning VDoplomacy (even from some moderators) has been able to make adjustments to the ratings system.

There's a meme circulating on social media now discussing an exercise by the US Navy during WW2. The meme has the image of a plane accompanying it with red dots indicating the location of locations the returning planes were taking the most gunfire. I don't know how to link the image here, nut the text is a telling example of what happens when people use static analysis, and fail to think through the entire issue before pushing through changes.

"During WWII, the Navy tried to determine where they needed to armor their aircraft to ensure they came back home. They ran an analysis of where planes had been shot up, and came up with this.

Obviously the places that needed to be up-armored are the wingtips, the central body, and the elevators. That’s where the planes were all getting shot up.

Abraham Wald, a statistician, disagreed. He thought they should better armor the nose area, engines, and mid-body. Which was crazy, of course. That’s not where the planes were getting shot.

Except Mr. Wald realized what the others didn’t. The planes were getting shot there too, but they weren’t making it home. What the Navy thought it had done was analyze where aircraft were suffering the most damage. What they had actually done was analyze where aircraft could suffer the most damage without catastrophic failure. All of the places that weren’t hit? Those planes had been shot there and crashed. They weren’t looking at the whole sample set, only the survivors."

The original VDip rating system wasn't perfect, chiefly because it was allowing for a lot of fraud, and cooked up games against straw man competition. A strength of schedule modifier was necessary. Maybe even a penalty for losing games was worth a look. However, the powers that be listened to the screaming mob and put massive penalties through for top ranked players losing, and huge incentives to other players to kill these players. In spite of several calls for caution in applying these modifiers, the system was changed to the mess we have now.

The style of play debate you guys are having is somewhat tangential to the ratings conversation. However, from my perspective the present ratings system is compounding the overall care bear attitude overcoming the site. It's too dangerous to lose, and the penalty for losing is not worth the risk of reward. So players generally take the easy way out, and we see too many premature draws. When this becomes a trend across the community the game doesn't function the way it's supposed to. Any socring or ratings system applied needs to stress the rewards for sucess/victory/solo while not penalizing people who play to win and fail with anything more than losing the game. Not points or rating system should apply a negative score to a player who puts in months of dedicated play. We apply no penalty to players who NMR, CD, or sabotage games, but we penalize players who are loyal to the community and the hobby at large. That kind of idiotic logic is what will ruin a gaming community.

Which comes around to my point I've been making for a very long time now - mainly that we shouldn'td have any points or ranking system at all. If people care about that sort of thing then the raw data of win/draw/survive/defeat on a players profile should be more than enough.

I can just remove the point system for a while so we can see how this will change the behavior on the site. It would set all (new) games to unrated, does not display PPSC and WTA and remove the points displayed behind each player.

I am really not in favor of removing the rating system.

We have seen, that the current vPoints have their flaws encouraging headhunting, discouraging play for high rated players and perhaps overrating success on very large variants. But do I believe that we can find a better solution than to remove the rating system completely. Mercy already made some good suggestions in that direction. And I will try to add further on this in the following. But first on the need of rating systems:

I for myself enjoy our many parts of our rating system and am convinced, that it does motivate many players to play more competitive in rated games. As a consequence I assume that rated games tend to have a higher quality than unrated ones. Though I admit that this is really hard to quantify and I can only speak on my own behavior in such games.

Just limiting the rating on the win/draw/survive/defeat statistics is insufficient on vdip, as the variance of variants and their required player numbers is too large. It is obviously much easier for an over-average player to get a high ranking with small variants than it is with very large variants. As we have already seen, that ratings do have an impact on the behavior many members of this community play the game and the consequence of restricting ourself to the win/draw/survived/defeat ratio would probably just make it even harder to get variants with many players going.

So that the vRating takes the player number into account is in my opinion a good feature per se, as it makes success in different variants comparable. Though if there is a consensus in the community that large variants tend to be overrated, we might want to apply the 'incremental way' suggested by Mercy or just add an overall weight factor that is smaller for large player variants. In any way we would have to set those adjustments by experience or does anybody know a better way to reasonably quantify the comparability between different variants?


If we want to discuss rating systems, we first have to find common ground on what the rating should achieve.

If we want a rating to just compare the overall success on the website, we could use a rating that just rewards wins and to a lesser degree draws and maybe survives, too, but does not punish you for performing not that well in a game. A positive consequence of this would be that high ranked players would be encouraged to keep playing so they are not overtaken by their competitors. On the other hand new members and those who do not want to spend that much time to the game might have no chance in competing and might therefor be discouraged to play or at least lose that motivation factor of a rating.

On the other hand if we want a rating that should compare the skill of players among themselves we need a measure that quickly converges to a value reflecting their current skill level. Taking the player's and other game members' rating into account and punishing objective bad performance, i.e. the final results, in a game is actually a good way of achieving this. Though of course players that had a good row of games and feel overrated might be discouraged to play to avoid being downrated again. In general as such a rating aims to converge to a certain value reflecting your current skill, the objective changes from increasing to just holding your rating at a certain point. So from the point you reached your "fitting rating" you should only progress with your skill increasing which is of course much less motiving on a psychological level if you ascended steep before. So we certainly lose the factor of encouraging play here on a general basis.

At this point I have two ideas I'd like to propose to tackle this issue and preserve a quick convergence and no punishment for small game numbers. Both approaches stress the comparison of the *current* state in a similar way:
My first approach is to add a slow decay / convergence with time to the average level of 1000 points. If players had not played on this site in a long time, less information is available on their current skill so it seems reasonable to assume them more average. High rated players are encouraged to keep playing to even hold their current skill.
A second way is a rating the temporarily resets. This way players are encouraged over and over again to play the game and everybody has the same chances to achieve a good rating at a certain point. Of course we do have a problem here with finding a time span large enough, so players with only a few games or with very large games with long running times can have the rating converging long enough.

As a member who does not play that many games I am more in favor of the skill approach the vRating takes, perhaps with modifications like the temporal decay that tackle the discouragement to play.


Concerning the headhunting: That is an effect that we should definitely try to avoid with a rating system.

One way is proposed by Mercy and the Ghost Rating. I am not familiar with the Ghost Rating, but just calculating an overall rating of the game which does not depend on how the other members perform in your game seems to be reasonable to tackle this. The consequence with the rating in draws is however a point, one should further think about to find a rating fitting to the large number of different variants.

Another much simpler solution that avoids that issue is to not taking player skill into account at all, but just player number. Of course this would be huge step away from the current spirit of all point systems and ratings we know so far, and goes more in the direction of the win/draw/survive/defeat statistics, as all games with similar member count are treated equally. And unfortunately it also has some serious drawbacks: Convergence only happens if there is something like a temporal decay as supposed above, when a players average points gain rate equally the decay rate. This does of course rate players with high number of games better, but at least it does therefore encourage members to keep playing. So perhaps this might serve as starting point for a good rating, too.


To conclude, if we decide to change the rating, I would at this point prefer a rating similar to the current vRating but with the following modifications:
- Applying weights to the scoring according to player number to reduce impact of a large variant win
- Adding some decay factor to encourage player to keep playing and defending there rating
- Adjust the scoring in a way similar to GhostRating to avoid headhunting
Of course some more thoughts would have to be put in each of this points and perhaps someone can convince with a completely different approach.

ubercacher16:
"Players will always want to draw with other players who they really enjoyed playing with even if that means forsaking an obvious solo route, you can't change that by getting rid of PPSC."
I actually think that this is far more likely to happen in WTA games. In my experience, players have become terrified of pushing for the solo and are shocked when you refuse to draw in the mid-game. In PPSC games, players always have a points-incentive to try the obvious solo route.

RUFFHAUS 8:
"However, from my perspective the present ratings system is compounding the overall care bear attitude overcoming the site. It's too dangerous to lose, and the penalty for losing is not worth the risk of reward. So players generally take the easy way out, and we see too many premature draws. When this becomes a trend across the community the game doesn't function the way it's supposed to. Any socring or ratings system applied needs to stress the rewards for sucess/victory/solo while not penalizing people who play to win and fail with anything more than losing the game. Not points or rating system should apply a negative score to a player who puts in months of dedicated play."
This is my analysis of what happened to the playstyle on webDip when they switched the default scoring system from PPSC to WTA. I mean it's striking, really. You can apply the same sentences word-for-word.

I don't have time for a more elaborate response, but want to say a few quick things.

Just as tobi, I think that it is best to look for a few improvements to vRating that everyone could agree are better than what we have currently. Throwing away points and ratings altogether seems like a bad idea to me; like I said before, I think it would lead to frustration if suddenly WTA and PPSC players became mixed in games where it was unclear what the objective was. Doing away for vRating for a while wouldn't be that bad necessarily but doing away with points would, in my opinion.

Let me now respond to the proposed changes of toni.

(1) "Applying weights to the scoring according to player number to reduce impact of a large variant win"
I do think large variants are slightly overrated, but I don't think this is the core of the issue. The problem is big wins in large variants. I don't mind people getting a significant rating boost if they 5-way draw in WWIV, but I have a problem with someone soloing in WWIV getting to the #1 spot instantly, especially if that would not have happened if he would have had a HIGHER rating prior to the solo. Hence my suggestion to take an incremental approach. I am glad that you agree that it works as a solution.

(2) "Adding some decay factor to encourage player to keep playing and defending there rating"
I agree that players 'hoarding' points is an issue. With GhostRating, players just drop off the list if they don't complete any new rated game within an X amount of time. They get re-entered whenever they do, and then their old rating is taken into account. I personally favor this solution.

(3) Adjust the scoring in a way similar to GhostRating to avoid headhunting
Yes that needs to happen. I do not think it is trivial at all to do so, though, but if people want me to do it, I can try to work something out. I already do have an idea for a solution, but I am afraid that it would be computationally infeasible since the computation time increases factorially with the number of players. So some kind of approximation that still works would have to be worked out.

Mmm that turned out to be decently elaborate after all.

tobi1:
"I am not familiar with the Ghost Rating"

I'm surprised! It's been around for ages. You can read up on it here:

https://sites.google.com/view/webdipinfo/ghost-ratings

https://sites.google.com/view/webdipinfo/ghost-ratings/ghost-ratings-explained

@JECE, I was making a point. That problem isn't caused by a particular rating system, but I thought that using that as an example would punctuate my point well, also I was in a hurry.

@ Mercy: Just some quick response to your reply.

(1) Weights vs. incremental approach:
I had another thought on your approach and believe I misunderstood your issue before. So am now I correct that your issue with the current rating of large variants is, that it is possible for low ranked player to have a higher rating after winning a game than a high ranked player? In that case I see the advantage of integrating and can agree on this being superior to weights.
Personally, I do not have a problem with 'ridiculously high' increases as I assume those to quickly oscillate back anyway if the rating is unjustified (and if the player continues to play). So I can live with anything here that is not too complex but reasonable. ;)

(2) Treatment of past results:
While I do believe that the GhostRating approach solves most of the issue and I would certainly live with it, there is still some issue, that it might encourage less risky behavior by high rated player. A player might for example tend to only play a few, small games to keep them in the rating but not risking to high losses. This might be a bigger issue here than on webdip as we have a larger set variants.

(3) Adjustments for headhunting:
I am curious to hear about your idea. Though I have to agree that factorial dependence might be a problem for a 50 players variant (even if a game of such a variant only finishes once in a few months).


@ JECE: Thanks for providing the links. I do have heard of the GhostRating. But since I do not play on webdip I had not bothered to explore it so far.

@Uber -

I'm glad we could have a productive conversation and help each other figure out what the other was trying to say. It's far too rare nowadays, especially on the internet, for people to be able to have a conversation that actually produces some sort of result instead of just slinging mud at each other. And your chess example was a very useful one to show exactly the point I was trying to get across!

That brings us to the "how" of bringing about that vision, since there are *many* people on the site, who DO view draws as an acceptable "win" condition, and as such, are changing the game without letting others know beforehand what they are getting into.

Obviously this is a mind-set change. As Ruffhaus pointed out, we see way too many premature draws simply because people don't either want to take the risk of going for the solo, or because they view the draw as acceptable. What you often see is a game where it's obvious years and years before the game ends, just how it's going to end. Those games are really just in the mid-game. We need to encourage people to take that next step, and really make an effort for the solo, even if it's unsuccessful.

I guess that's where the ratings system could come in. Encourage the solo, heavily perhaps? Minimize rewards for draws? I'm not sure what the answer is, but whatever the answer is, it'll have some heavy lifting to do in order to reverse the complacent mindset that seems to have set in here where people are fine just drawing in the mid-game and calling it a "good game".

tobi1: Not sure why, but I thought you were a veteran of webDip, or even phpDip.

Couldn't agree more drano.

I didn't want to post on here earlier because the game was gunboat, but thank you for the shout-out. The game has finished now, but I did wind up with what I'm pretty sure is the largest army and highest SC count ever seen on vDip before the game ended. Sadly, I wasn't able to get a solo (probably because I had NMRed several times that game and failed to take advantage of several major openings early enough), though I do consider a solo on the Divided States map to be a possible feat given the right opportunities and the right strategy (by my guess, that involves focusing on land in the early to mid-game to grab as much as possible before it stalemates, then building up a navy, preferably on both sides of the PNC to NAS or NWP chokepoints).

As for the ratings debate on here, I see how crazy it is that a single game can give such a large boost. My own ranking shot up 400 points from this one match (I'm apparently 15th on the site now).

@Technostar: Wow! 50 states, down to 3. Impressively played.

That is impressive indeed, Technostar. Congratulations. :-)

On Divided States, I've managed to identify two crucial choke-points: The Alaska coast and the Panama Canal. For Panama Canal, only 7 fleets are needed to lock it from the West (more from the east), and only 3 fleets, plus a moderate land presence, are needed to lock the Alaska Coast from either direction. This of course depends on a rapid domination of the entire coastline on one half, most likely from the west. Therefore, I suppose this variant has a slight inherent bias towards nations situated in the west, who can relatively easily and quickly put themselves in a very easily defend-able position with just 10 fleets.