@goldfinger: how the fuck can lando/sarg/jemeyer be lower rated than jmo, when they all beat jmo at the only category which jmo plays. I recall my example with the languages earlier. It's absurd you consider a player better who is beaten in the only category he plays by guys you consider worse. You can't compare jmo to lando/sarg/jmeyer in the current system, yet you do it.

I know Lando is probably losing in world games, etc, I'm aware of that, and that's the point this whole thread started from, that many player don't play other types because he is afraid of losing points. Lando plays and he is punished because of thet even though he beats the crap out of many players you currently consider better than him.

@goldfinger: the whole point of the new average would be that your other types wouldn't drag your rating. The currently counted subcategories are examples for me when I do counting. But it's simply stupid I beat a player when we play the only category he plays, yet I'm rated behind him just because I played different types as well.

As I also stated in an example it can happen you improve in a sub-category yet you lose rating (not ranking) in overall. This is exactly what happens to Lando. He can improve in fullpress or gunboat, but he will lose points overall at the same time.

I don't get why uclabb think I play only 1 game, I make a 3way draw, my fullpress rating improves, I get higher in the category, yet I can get lower in overall rating.

Do you think just because I made a 3 way with better fullpressers I should get lower in overall? How the fuck is that? I beat better players and I get lower overall. And you guys say that normal?

@roka- I'm not saying that at all. You are putting words in my mouth. What I am saying is that using one number can never be perfect. So pointing to a problem without context is meaningless. I think that your proposed solution has many more problems than the current system, and I have tried to point out some of them.

Again, the problem you are pointing out is, I think, not a big one. It seems to me that strong full press players are perfectly capable of playing gunboat and live games at a high level. I really don't think that really anyone avoids such games for fear of losing GR. On the flip side, I do think it is possible that gunboat players who care about their rank avoid full press games. But, to be honest, if a gunboat player can't play full press at a level comparable to that they can play gunboat, then their rating should reflect that. I admit, there is a built-in full press > gunboat bias here, but there should be. The game is called diplomacy.

@uclabb: I'm at 118pts. in FP category, I play a 150pts. average game, risking 2/35*118= 6.74 pts. have a 3-way, meaning I will get [ 7* 2/35*150 ] / 3 = 20 - 6.74 = 13.26 GR points right? So I will improve 13 in FT category (10 places in ranking)

Meanwhile in overall:

I'm risking 378 * 2/35 = 21.6 pts. The game can be 150 pts. average by overall GR to, meaning I get 20-21.6 = MINUS 1.6 pts.

Resulting I beat better fullpressers, waaay better, I improve in my category, I don't play anymore games in that month, and at the and I lose GR rating and ranking. So why the fuck would I play fullpress then?

The only problem you pointed out is you fear players haven't played enough but I disagree with that, since they haven't played enough in ALL categories but they played enough in the category they are good at. Meaning your complaint simpy doesn't stand. The more they play the closer they get to their average.

As my example shows, the more I play, even beating better players, the farer I get. That's absurd.

"many player don't play other types because he is afraid of losing points."

That is absurd. I don't play other types because I don't enjoy playing them, not because I'd lose. Put me in a live game and I'd kick all but the top 50 players' ass, but I just don't like the amount of stress live games put on you, nor do I enjoy typing for 3 solid hours at least.

but as for your main point, do you really care that much about rankings to install a whole new system? The current one isn't bad. Besides, as with all ratings systems, its just a generalization. The Rangers were supposed to be the best team in baseball, and where are they now?

What you are saying is the same thing over and over.

" I disagree with that, since they haven't played enough in ALL categories but they played enough in the category they are good at. "

This is a non argument. You are punishing people who don't play variants of diplomacy and giving a disproportionate benefit to players who play more games Both of those things are true.

It's like you learn martial arts, and the guy who is a black belt Karate Kid but doesn't do aikido will be considered better then the guy who is Black belt at aikido and beats the crap out of green belt karate guys with only having a yellow belt in karate.

And after he beaten those karate guys you detoriate his martial arts rating and still consider that black belt Karate Kid the GOD and best ever at martial arts. Poor Aikido Kid will never ever learn Karate anymore, when people like you say him well, it was nice to beat those green belt guys, so we degrade you back to white belt from yellow :D :D

Roka, considering your comments regarding the low pot-high GR game we played, I'm surprised that you are making such a big fuss about this.

The GR, as we know it, has historical significance to this site. I'd be skeptical if a new 'overall' rating system was ever adopted. I also have some issues with GR but, once you accept it for what it is - a crude point system thy rewards you for beating good players, you can be very happy with it. It's an unofficial scoring system that's been adopted by this community and, while it has its flaws, it serves its purpose of providing goals, challenges and bragging rights.

Because of the small amount of games that diplomacy players play, I doubt that a reliable system can be created to show people's 'true skill.'

Beyond that, I'm not sure hat you're trying to accomplish with this new system? You want a better way to compare the skill of Bartdogg and split diplomat? I think you're creating an issue where there isn't one.

Yes I'm punishing them for playing less variant because I'm looking the best layer on the WEBSITE, not in a board game called classic map fullpress diplomacy. This website has more variants and types, so if you are looking for the best that guy should play more variants and types, since he is the best on the fuckin WEBSITE not in a board game.

Also I can punish those who play less, once again, the guy who plays more on the WEBSITE deserves to be considered better against another player with whome his true GR equals. Why would a guy with 10 games be better (with a true GR of 999) than a player with 100 games (with a true GR also 999)

I think it's a valid punishment.

Yonni, you are using current overall GR to define who is BEST on this WEBSITE. I'm only arguing with that assumption stating the simple fact that overall GR is all but showing who is best on the WEBSITE.

All you wrote down I totally agree when it comes to category GR, but I disagree when it comes to overall GR. Overall GR doesn't show who is best on the website, neither "rewards you if you beat better players" That simply WROOONG. I just wrote an example to that one, it actually punishes if I beat better players.

@roka- It sounds like we have reduced are arguments down to axiomatic differences, so I guess we have done our job. I disagree with nearly everything you said in your post that replied to mine. Live games and gunboats were added as features long after I joined the site. At its core, this site is about the game diplomacy. vdip is the variant sister site.

And the best player on the site can have played only 10 games. When Edi Birsan played on this site for a while, he was the best player on the site, despite his low number of games. GR didn't immediately reflect this (how could it?) but it was getting there. But that is actually one of the core beliefs of GR: that total number of games should matter as little as possible without sacrificing stability in rating. If you just care abut volume, just look at points.

I really have to sleep it's friggin 4am, and I'm arguing with people who are defending a system which is not just wrong but against the site's interest. Ridiculous.

This system punishes you for playing more (which is against site interest), this system punishes you if you play other then fullpress classic (which is against site interest), this system discriminates you because you play other variants and make people think you are a lower class person (which is against site interest)

Why the holy fuck does this site have variants and game types if you guys want to evaluate players by a system which doesn't tkae those options into account.

It's like let's find who's the most clever person on the World, but let's exclude blacks, asians and women first. becuse they are boring, or they are not the ruling elite, or whatever. :) Well you can do that, just please don't call the guy the most clever HUMAN after.

I went to sleep. bye

GR is not part of the site. It is external to the site. Points are the thing there to encourage people to play more games and to not quit games early

Lando's situation is pretty odd, I am not totally confident but it seems pretty similar to an example of Simpson's paradox.
http://en.wikipedia.org/wiki/Simpson's_paradox

Maybe I have the wrong idea about GR, but I thought it was an attempt to quantify the skill of a player at diplomacy. The only data they had to do this was wins/draws/loses. Each game of Diplomacy is a little crazy, and a little randomness sneaks its way into the results, that is why there needs to be a relatively large sample size of games to get an accurate result.

Generally a large sample size is better, more games tends to filter the noise out of the data. When the GR is categorized is is basically ignoring data. You gb GR is ignoring all of your fp games, and thus losing out on a lot of data. Your overall GR takes into account ALL of your games, no data is excluded. In this way your overall GR gives you a more accurate result of how good you are at Diplomacy(assuming diplomacy is defined by all the options of games that are on the site). If you are great at gunboat but suck at fp, then according to your overall GR you are somewhat-ok at Diplomacy.

Regarding the whole more games=higher rating: I think that to a certain extent more games does tend to increase your rating, but that is only because it takes more games for the algorithm to get enough data to give you that higher rating.
If we assume your true GR is 200 and you start at 100, we can be pretty confident that your first 5 games will generally raise your GR. It naturally takes time to raise it to a state of equilibrium. After a substantial amount of games, more games don't necessarily help raise your rating though.

@uclabb:

what the difference is between us ideologically is you say diplomacy is a full press game.

You'd be right, if this site was about playing diplomacy. but this site is about playing WEBdiplomacy, which means it's not face2face, not live, and so on. A lot more changes. Still you want to honor those who are good at beating up skilled players in few games, those who play fullpress, and so on.

It's like honoring a player in need-for-speed just because he can drive in ral life, or honoring someone in counter-strike because he can actually shoot in real life.

This is a network game ONLINE. This has nothing to do with the live board game skills called Diplomacy. I undertand you joined when it was much more alike, but now it isn't. Time has passed, the site is developing, player are changing thanks to online features, yet you still want to use old rating methods. Guys like you are those who don't let the world develop, because they are so conservative they don't want to see any changes, even though the world itself has already passed by them.

What you should understand, the guy who ought to be the best player on this website, at this online game, is not the same person who should be best at the board game called diplomacy which is played live face2face.

This is a website, which wants more players, bigger community, more game, wider variety to be played. To achieve that goal the rewarding system should support that goal, not obstruct it.

But it also shouldn't demean those players who only play non live classic fullpress by making their games become so outweighted due to sheer volume of live gunboat games. Any cumulative system does that. Only a percentage or average based system doesn't punish the "slow game player".

"What you should understand, the guy who ought to be the best player on this website, at this online game, is not the same person who should be best at the board game called diplomacy which is played live face2face."

Have you played a F2F player on this site rokakoma? They regularly beat the shit out of the rest of us. TheWizard is an example of an excellent player F2F and on this site. This site is about diplomacy, no matter how much you may disagree with that assessment.

The ratings system on this site is fine. If you care to tell why or why not a person is highly ranked, look at their game history and that'll tell you why. Like I just looked and saw you play mostly gunboat or live games, which is probably the reason for the disconnect between you and the full press players on the site.

@Draug: my suggestion is to creat that damn list, ask Alderian kindly to make it, let's see who it looks like, how it changes by time (he can create it for the past months) and then we can discuss.

According to the suggestion there are 8 categories, personally I play 3 of them (normal FP, live gb, normal gb, all of which is WTA)

Meaning many other player don't play 8 categories, so those ones who are good are playing 1 will be hardly outweighted, since all the other 6-7 categoies they will be at 100 pts. It's hard to outweight someone having 150-160 GR points in 1 category.

To outweight someone with 150 GR (giving 150^2+70k = 95k sum of squares) you need to be at least 120 pts. GR in at least 4(!) categories. (3 gives 93k sum of squares)

So it's really not that easy, and outweighting even better players gets even harder.

@goldfinger, 70% of the games are gunboats and you guys totally ignore that fact and keep gunboaters as a lower class people. Just because you don't respect gunboaters it doesn't mean they don't exist.

On the other hand I think fullpressers has like zero strategic skills, I don't like the way they play, I always have to describe them what will happen in 3-4 years on the map.

All that said, I don't see why we should treat the 2 types of players equally and motivate gunboaters to player more fullpress and viceversa. I would love to see MM or zultar playing more gunboats, as Split and I would gladly play more fullpress. But currently, the current rating system discourages this effort. I'd like to create a rating which engourages you and me, to play those types which we rarely play.

I don't see what the heck is the problem with that.

@Roka - Equal players play for 2 months. Both win all.their games. The live player plays 50 games in that time. The non-live plays 5 consecutive ones. Thevlive player gets about 10x the GR boost of the non-live. It is a mathematical reality.

That's the thing though rokakoma, in full press you can't think 3-4 years ahead. Anything can change any turn. The beauty of talking. Gunboat and full press are simply different beasts and you can't go from one to the other with the same preconceptions.

I honestly doubt that the rating system discourages people from trying out the other variants because 1) They most likely don't care THAT much about their ratings and 2) They probably don't play the other variants simply because they don't enjoy them. Theoretically, you could make the argument that moving around variants would lower GR and people would be adverse to it, but in practice I simply don't see it.

@Draug that's not true, we have discussed it many time. The more games just takes you closer to your true GR rating, but that's all. Just because you play more you don't get more GR points once you are around +-10% from your true GR. That's only an issue for the newbies, anyone who has played more than 20-30 games in a category doesn't matter anymore.

As uclabb said yesterday, yes this average punishes less games played, but that only an issue until you played less then 20 in a category. From month to month it doesn't matter. Just because I play 50 gunboats. Look at SplitDiplomat he has played 300 games in the last 5 months, and his GR didn't even moved more than 3%

@Roka - First, please prove that stat (70%) and second, how many of those games are live versus non-live. My gut tells me those gunboats are mostly live where as the non-live games are mostly full press. And considering how many live games can be played in the time it takes to play a non-live, even if your 70% turns out remotely accurate (I highly doubt it) it isn't a fair comparison do to the time and energy required for full press non-live games to play.

@goldfinger, even if you are right on every single argument you say, please tell me, what the heck is your problem with seeing a list as I proposed which takes all types of games into account and places the one equally skilled player ahead of the otherone if he plays more types. What's wrong with that?

Also I'm playing a fullpress now, with my ally we planned 3 years ahead and it just works perfectly and everything went as I predicted. Just because YOU think it doesn't work in fullpress I have to say you are wrong it does work. People there are just not good enough to see that.

BUT, we are not here to argue about the beauty of fullpress we are here to discuss what's your and others' problem with such a list, which in my view would show who is the player on this internet website who utilizes the most type of games the best possible way.

I don't care about how you call that list, I'd like to see a list where diversity and wider spectrum of skills are rewarded, where someone, like Lando or Sarg, is not punished against players who play only 1 category even though they both beat that guy in that category.

Please tell me, why do you place ahead (because you do when you say the current list is good) a player who plays full press classic only against those players who play more types and can beat your player in his only category.

How can you think that he is a better player? Because he is not. The only way you can compare them, his a worse player, yet you place him way ahead.

@Draug: i said that by heart, Vaftrudner had stats about that and I just remembered saying me. I will try to dig the data.

@Roka - Please clarify which GR you reference: Rank (position relative to others) or rating (absolute number). I disagree with your analysis concerning rating cause it is gameable by playing all noobs and never losing.

I am on my phone so can't do this but I bet if you lpoked at the average gunbpater GR and the average Fullpress GR (top 100 players in both cases) you will find the gunboater has a significantly higher rating score due to the sheer volume of games they play live.

Okay I remembered wrong:

According to http://www.draugnar.com/VaftStats/

There were (all on classic map)
3025 live gunboats
4080 live regular
4698 gunboats
15759 regular