Ctp Pbem Rankings 12th April 2002

[J Bytheway]

Originally posted by blackice
As you can see the odds of that happening beyond a two player game are extremely rare. you put 10% of your points into the pot and gain 8% or less.

take the 3 player game example - players A, B and C in that order with scores a, b and c then the gain to C's score is

0.13(0.1a+0.1b+0.1c)-0.1c
=0.013a+0.013b-0.087c

So this would occur only if

0.087c < 0.013a+0.013b

Let d be the mean of a and b, then this is when:

0.087c < 0.026d
c < 0.30d

i.e., when c is less than about a third of a and b, or just 3 games wagers worth. I think that this is too severe a problem, but I admit it is less so as the games get larger.

Mine and a pile of others too. this was the biggest problem with the old system. We are just waiting for quinn to post this senerio so we can move on.

Now he has done so, and I presume your concern is the large gain of Darth (larger than FG or Kralj), but a 12/13 point difference is very severe under this system and so Darths 3rd place, especially above Bird shows that his score is vastly off and must be increased. I would guess that in this case it would approach 17 or so. Of course if he were in other games where a lower score would be more appropriate the equilibrium would shift.

Lung senerio is the fact he has one game and the opponent is lower ranked. If you have first place you have to work to keep it. Lung will also win points when he wins this game. no system should allow a 1st ranked player to play but one game against a weaker opponent and remain in the top spot while others have several and against better opponents.
The Klair factor will not happen in this new system.

True, but at the expense of placing an arbitrary upper limit on rankings.

Correct be if I am wrong but that depends on his opponents winning. if the opponents are losing games at a faster rate this player will continue to drop in the rankings by large amounts.

Well indeed - but losing to players who are in turn losing many games should warrant a fast drop in your score, should it not?

This also would depend on how many games the player is playing. If only one game it could put them down huge. It also depends on the other games they are in. If they are winning other games they are gaining less than they are losing. Back to the 2 wins mean less than one loss.

And so 2 wins should mean less than one loss in this situation.

The average person does not have the calculator or the knowlegde to make these formulas work.

I find that hard to believe but (at the risk of sounding immodest) I am in the habit of overestimating others abilities in that area.

That is a nessessity in order to maximize ones wins with the old system.

Are you saying that an understanding of the formulae is a necessity?

That was the other problem because of the huge wins vs loss difference. With the old system you would want to play a game with less or no players way down the rankings to maximize your win vs loss on the ranking board.

You only need avoid such game sif you are worried about losing them. The problem, of course, is that CTP games are determined as much by luck as skill (back to the terrain dependance again).

The new system does the same thing but with less of a hit both ways, winning or losing. This is more friendly to PBEM as the chances of playing lower ranked players are great. The fact less damage is done promotes games with lower ranked players. This is needed without the need for rules. This also eases the "bad land" scenario that is a huge factor. You are playing lessor ranked opponents you have tundra they have rivers and mountains. Like gary said we should take the time each and every game to balance the land factors.
Any volunteers?

Again, the only solution I can offer is levelling the graph in the fashion described above.

I think I like the differential system more simply because I can see (on the tip of my brain, as it were) how it is nice for mathematical analysis, but I agree it has flaws in the CTP context - due to the false assumptions of fair games and independant results upon which it is based.

I feel that the wagering system has more problems that are as yet undiscovered simply because the system has not been used as much yet. It is much more difficult for me to see exactly what is happening in the system mathematically because, believe it or not, it is much more complex than the differential system from that perspective since it depends on absolute scores rather than just relative ones (I thought it might be sufficient to consider the ratios, in which case it would in fact be a differential system conjugated through an exponential bijection (which would have been moderately humourous, though perhaps only to a mathematician ), but upon further examination that was not in fact the case).

I don't suppose someone has time to do scoring via all these different methods so we can compare them in a more holistic fashion? (And no, I'm not volunteering, although I might be able to write a program to do it if someone else feeds in the numbers...).

[blackice]

1. Darth Viper 10.70 (D) +0.90 = 11.60 +1.91 = 13.51
2. Nuke Boy 15.00 (N) -0.07 = 14.93 +0.97 = 15.80
3. Faded Glory 22.30 (F) +0.38 = 22.68 -0.11 = 22.57
4. Kralj Matjaz 23.30 (K) +0.64 = 23.94 -0.88 = 23.06
5. Bird Man 27.90 (B) -0.97 = 26.93 -0.96 = 25.97 +0.13 =26.00 15x the loss for the win

Now if the next game Birdy lost again, the example above.Then proceeded to win in another game with the same people and he won he would get 15x less points than he lost. For One 9th round ranking playing three games. (-.97 + -.97 vs +.13) Losses mean more than wins no matter what way you look at it...

Now don't give that he lost to lower ranking players stuff of course he did. He also lost to terrain, goods, diplomacy and barbarians. The lower ranked person should not be greatly awarded for his bad luck! It may not have been skill in other words.

While Birdy is dropping fast the other lower ranked players have made little if no gains at all. They move up the ladder very slowly again if at all. The only player that moves a lot is the losser. If the higher ranked plays the lower ranks the gains are extremely small but the loss is huge! Not balanced for CTP, tennis and chess maybe but those are ranked at game end and played on a head to head even playing field bases. I could see this system would work if it was at end game maybe. CTP is not a level playing field. Because of that fact alone the win/loss ratio is out of skew.

That I believe is the prime reason no one could catch Klair. While this equalibrium is had, the odds and the points went down also. In time in the same games there would be no way to make up the losses the win/loss ratio was too great. The bottom line was it was better to take a one time hit. Quitting the game and getting into a new one and take your chances there. But do not remain in the old game where chances are you would gain very little points after the equalibrium was attained or obviously more so before for that matter.

1. TheBirdMan 24.48 10% wagered =2.4 - 4.39 = 1.99 new ranking = 26.47

In the new system 3 x 9th turn rankings gets Birdy moving close to his old ranking. The other players while winning their other games against Birdy move up the ladder too because they won. The pot gets larger and the winnings get better. That is a huge difference you have a better shot at gaining rankings again as the games go on. This is the built in "stay in the game" factor. You win more as the game goes on everyone does. Yes an equalibrimum is met in the new system also it is just not downward as much more balanced because of win/loss ratio.

In this case Darth would move up for his win and some for his third. This system promotes winning games and awards for winning and staying in games. Something we could use here without a rule added. Less rules the better and since this system does it, no need for a rule...

At the same time the rankings are constantly moving up and down but in equal proportions. For first and last place players it is a lot sometimes. For second and maybe third move either way mostly second sometimes third. Could even be fourth if they were ranked low enough and the pot was large enough. Fifth not likely but at the same time thay aren't creamed either.

The huge difference is what the winner gets the lions share. This balances the system out and makes for a better system when dealing with things like land, barbarians, goods and diplomacy in CTP. So yes you do get more points from the higher ranked player if he losses. The beauty is you get them primarly from the higher ranked player!

That is the one biggest difference in the two systems. The other is the Klair factor. The new system takes in account the terrain, barbarians, diplomacy by not slamming the loser. It promotes games with lower ranked people. And the winner baby is the top dog! Not the loser is dead. It is more balanced in it's win vs. loss ratio with the winner winning the lions share and the rest picking up the scraps. It is easy to calculate meaning I can do it on my windows calculator. It has worked so far without one glitch. It does not need rules to make it work. And it was designed as a CTP Ladder Ranking System. Tried tested and true in GL for several years.

It was adapted here buy lessening the wager to reflect the fact the games are ranked on 9th turns not end game. All in all a win win scenario the winner wins the loser losses. Just not the huge extremes like the old system.

Lets address Lung now. Lung match the game he is in Lung can gain points for winning as shown below. If the game is down to 3 players he would get 1.05 for winning a 9th rank turn. If it is still a four player he would still win points same with five players. I think someone made an error when they claimed that he could not gain. That is if he is winning of course.


LUNG.......................30.8350 10% wagered = 3.0 -4.05 3 player 1.05
Mobius.....................23.3120 10% wagered = 2.3
Faded.Glory................22.2950 10% wagered = 2.2

TheBirdman could potentially take over the lead, but it really depends on WHEN results come in - At the top if you come last in a 7 or 8 player your score will drop by 3 points, and a win in a 4 player will gain you maybe an extra 0.7-1 points.... It gets harder to break clear of the pack because you stake more points on each game.... Thats why I love this system so much...It seems fair, consistent success is the only way to take top dog spot..

I agree with Gavrushka here one hundred per cent success wins the world. Unlike the last system where lossing was the difference overall. In the new system winning is the name of the game and you get the lions share of the pot.

Quinn you said Darth should be ahead of Birdy because of skill level. In chess maybe but factor in wins, losses, terrain, barbarians, goods, diplomacy in your thinking. It is not just a head to head game on a flat surface with no other factors. I mean even in tennis you have the court factor, clay etc or injuries.

In chess you may have a headache or a distraction, scratch that you can have distractions in any game or a headache. Way more factors in CTP to include.

Bottom line it would be nice to have a system which totally relies on skill but it does not suit PBEM CTP very well.

The new system does a better job at compensating for the varialbles with a better win/loss ratio. And less rules to make it work

[blackice]

Of course if he were in other games where a lower score would be more appropriate the equilibrium would shift.

Maybe far to often in the past this was not the case. People would rather take the one time hit than stay in the game to lose points. In the new system third is third and more points are gained by all the longer the game went on...

True, but at the expense of placing an arbitrary upper limit on rankings

Not that I see the Lung problem was wrong.

Well indeed - but losing to players who are in turn losing many games should warrant a fast drop in your score, should it not?

Yes and the new system does that with more balance when you lose with the old system it takes forever to gain back what you lost. The main reason is your wins are generally small but your loses are generally huge.

The new system you win you take the lions share of the pot. You lose yes you take a good hit but because the way the system works your gains on your wins have you going right back up again. This is true especially for weaker players who may see 5th more than first.

Most games see weaker players ganging up on the top dog here. If the ranking is the determaning factor. Then it only makes sense right of the get go for people go after top ranked all four is smart. That way the top dog feeds the rest points. Same in the new system but with less of an impact so it is not such a big factor.
More balance between wins and loses is the reason and the shared pot.

And so 2 wins should mean less than one loss in this situation.

I spent all day on this doing calculations. In the old system as per the example above Birdy in no time could need 15 x as many wins to regain what he lost in two games. The main reason is the ranked players in the old system go up small and he goes down large. The new system worked out to about 3x more balance there. I think this may be the root of the problem with the old system.

I find that hard to believe

Believe it just ask Gavrushka or me I found an extremely nice calculator online that did the work for me today.

Are you saying that an understanding of the formulae is a necessity?

Absolutely yes if you play for ranking you use every tool at hand to assure success. Go back to

CTP games are determined as much by luck as skill

Too true at least by contolling who you play in the old system you had contol over that. Not land, barbarians, goods or diplomacy. The new system balances those facts out.

[TheBirdMan]

Don't think many actually do cares who they play with/against - as long as they play their turns .

Fx. I would welcome ANY players IF they promised to play their turns without any un-reasonable delays and kept their promises .

[Kralj Matjaz]

Fictional example of Gavrushka/Blackice system:

LUNG.....30.0 10% wagered 3.0 = 30.0-3.0+2,898=29,898
Mobius...12.0 10% wagered 1.2 = 12.0-1.2+1.827=12.627
Faded....11.0 10% wagered 1.1 = 11.0-1.1+1.071=10.971
Kralj......10.0 10% wagered 1.0 = 10.0-1.0+0.504=9.504

Obviously Lung despite wining a game lose and against more lower ranked players he would lose more! Such system is nonsense!

[TheBirdMan]

No matter if we use a system from Quinns/Solver or from Gavrushka/Blackice, there is only ONE solution.

Such a game should be an unrated game, 'couse IF the difference between the best and the 2nd best player is that big, this is more like a traininggame than anything else.

[blackice]

Kralj let's keep it real here ok?

The lowest ranked, active member on the board is 12.000+ It just gets even higher after that. The odds of two more people dropping to that score are very high indeed. They all would have to be in last place in every single game they are in for quite some time. Then Lung would have to say "oh gee if I get into a game with these guys I will lose points" I think I will start one with them. You have better odds playing a lottery! So remote, your example is utter non sense!

LUNG 30.0 10% wagered 3.0 = 2.9 -.01 29.99 2.9 2.9 29.99 2.9 3.0 = +.1
Mobius 12.0 10% wagered 1.2 = 1.8 +.6 12.60 1.2 1.8 13.2 1.3 1.9
Faded 11.0 10% wagered 1.1 = 1.1 +.0 11.00 1.1 1.1 11.00 1.1 1.1
Kralj 10.0 10% wagered 1.0 = 0.5 -.5 9.50 .95 0.5 9.00 .90 0.50

As you can see birdy will gain points sooner or later. The longer the games goes on like this the more points he gets. Every other player in this game if playing other games would have to lose every single game for Birdy to lose points. As you can see within three turns of 9 ranking Birdy gains. Birdy would continue to gain also.

Now lets compare the other system as stated above if Birdy lost once to these guys it would take 15x turn 9 rankings to make up the difference. 3x or 15x you choose beside this scenario would be sooooooo remote again your better off buying a lotto ticket and winning

You can see kralj not likely Birdy would play this game keep it real ok?

[quinns]

"Keep it real???" Blackice, Kralj is just stating his opinion on the matter, just as you have in great volumes. No need to chastise him for stating his thoughts. And it isn't "birdy" or "tweety", it is TheBirdMan.

I think I see the main difference and problem between the GameLeague based rating system and the Differential based rating system. The two systems are really measuring two different things. The GameLeague system rewards those who play many games, and the rating number itself doesn't really represent much (quote "a number is a number") except that higher is better, that is it. The Differential based system does not reward players who play many games, and the rating is meant to represent a player's ability, not just some arbitrary number that floats without meaning.

That is: 12 to 14 is a "D" player.
14 to 16 is a "C" player.
16 to 18 is a "B" player.
18 to 20 is an "A" player.
20 to 22 is an "Expert"
22 to 24 is a "Master"
above 24 is a "Grand Master.

I know, Blackice, that this is not your intention with the ratings and rankings. Your intention is to promote many games and reward those that play alot, and just give points for playing, period (quote "it's a win /win situation"), and not worry about the meaning of the rating itself. My feeling is that we should not reward those who play many games, but have a rating that is somewhat meaningful in itself. But these are all just opinionated differences.

I think we have discussed this in great length (ad nauseam ). I'm not sure if everyone else is really interested in which system we use, as long as it is updated regularly. Maybe we should have our vote now?

[blackice]

"Keep it real???" Blackice, Kralj is just stating his opinion on the matter, just as you have in great volumes. No need to chastise him for stating his thoughts. And it isn't "birdy" or "tweety", it is TheBirdMan.

The whole idea was to give a realistic assessment not "an unrealistic what if" especially when "what if" does not exsist and not likely it ever will.

You keep it real that was the whole point. Opinions are great let's just keep it real ok?

I am sure if it bothered Birdy he would speak up for himself quinn what do you think? I call him Birdy in our emails I have yet to hear a complaint. Birdy does it bother you to have me shorten your online game name like that? You can call me Black if you like I care not it is an online game name. Heck I guy called me blacklace the other day
I will do as Birdy request's though if he does meantion it

The Differential based system does not reward players who play many games, and the rating is meant to represent a player's ability, not just some arbitrary number that floats without meaning.

You are right the old system awards for one game, Klair comes to mind... Look Quinn just like chess, tennis or any other game the more you play the more you move in the rankings. This is standard I see no problem with that what so ever, the new system also awards for one game, You will just not move up the rankings that fast. That is just like any other rating system and as it should be. Floats without meaning again you demean the winners of games here. Thier rankings are the way they are because they have won! Or they have lost either way it is quite an insult to say to them their hard work and good play is "meaningless"... The ranking represent the wins and losers here not just the losers.

meant to represent a player's ability

So does the new system but it also takes in all the other factors the old one did not. As J. Said CTP is a huge part luck not skill, land, barbaians, goods, diplomacy and ruins play a large part of the game. Not just skill you are more than aware of that. Your last PBEM with Birdy you had better land he lost does that really mean your skill level was far supperior? A system based solely on "skill" is extremely misleading and does not represent CTP very well.

Again this is not chess...most if not all the top ranked players online restart games to get a balanced even playing field of land. You know this too, if it was just skill there would be no reason to do that would there? I mean Yoda, paulino1 filix etc are very skilled players. That still does not take into account ruins, barbarians and goods displacement or the equal distribution of land but it at least helps the land somewhat.

Why would the tourney players make even maps and even goods distribution if it was only "skill" that counted you know that too. Why would they elliminate ruins and barbarians if it was just "skill" that mattered.
Simple they are tring to elliminate as many as the "luck" factors built into CTP to find out who is more skillfull.

We could do that here too and play without all those things personally it would get boring. Those things about CTP add flavor and chance. Those things give a lessor ranked player the opportunity to beat a more "skilled" player. Cool that is the way the game was designed I like it you?

Because if all that we need a system that represents all those things not just skill...The new system does just that and does it well...The old does not, it is not "skill" to have better land, 5 tech lead because of runs and many more goods than your opponent it is luck plain and simple.

I know, Blackice, that this is not your intention with the ratings and rankings. Your intention is to promote many games and reward those that play alot, and just give points for playing, period (quote "it's a win /win situation"), and not worry about the meaning of the rating
itself.

No not at all like any other system including chess or tennis the more ranked games you play the more rankings you get. I find this point here rather bland any system of rankings anyone playing more has more movement. Take Claude F. Bloodgood III he was in over 2000 chess games at one time. His rankings soared they claimed it was a flaw in the system, why should not have been ranked for all those games? Silly really. The man was in jail and had the time on his side so to speak

I would like to point out very few people have 25 games on the go not much time for that is there? On top of that it does not garantee they will win them all. To me this final point is mute. You can play in one game and move up the rankings quite well.

My feeling is that we should not reward those who play many games, but have a rating that is somewhat meaningful in itself. But these are all just opinionated differences.

Why? even in the old system if you played a lot of games and won not lost you would gain in the rankings be it slowly... This new system just elliminates the Klair issue which annoyed a lot of players to no end...I can see it now people drop all games but the one they are winning to maximize thier gains on the rankings. We do not need more people dropping games and or playing less than they can because of a ranking system and new rules. The only reason people what to put a rule in for more games as manditory is because of the old ranking system. the new system needs no such rule and because the way it is you gain more points staying in a game than dropping out. More win/win for the league and the players if you ask me. On top of that LESS RULES

We need a system that promotes games and promotes staying in those games. The old one does none of that without adding rules that could potentially hurt the league here. The new system on the other hand does all those things and more "without new rules" The new system is meaning full I take it to you losing is more important than winning? That is what the old system was based on losing not winning. I find it a far more meaningful systen that promotes winning and adjust's for the CTP variables the new system does just that as I showed in the above example.

Again this is not chess it is not a flat board and CTP has way more variables. Just because Paulino1 may be a good player he still losses to an opponent with better land or better goods or more luck with ruins. Luck is a factor the old system totally ignores those "Facts" Next time you play online trade techs from ruins to balance the game out. By doing so you get more "skill" involved still not totally because of goods, diplomacy, barbarians and land.

The new system is far more balanced for win/loss ratio this alone makes it a better system. As always quinn your opinions are valued as I hope mine are too? Yes I agree it must be maintained regularly. I was going to do it as the games came in. I hate things sitting around not finished. I program to do it automatically would help.

I am game to set up the vote but we need to refine the pro's and cons of both system in an easy to read format. plain simple and to the point. Would you like to work on that with me?

[jpww]

Is anyone going to update the ratings and rankings sometime soon, or what?

[blackice]

Once we agree on a system yes obviously JP

[quinns]

Originally posted by blackice
Look Quinn just like chess, tennis or any other game the more you play the more you move in the rankings.

This is the heart of your whole misunderstanding of the differential (old) rating system. Your statement is completely false.

Let's take the "chess" example. Gary Kasparov plays maybe 50 to 60 "rated" matches per YEAR, yet he is considered by most to be the highest rated chess player in the world. There are chess players that play perhaps 50 rated games per DAY! AND THEY WIN ALL OF THEIR GAMES BECAUSE THEY ONLY PLAY PLAYERS MUCH WEAKER THAN THEY. You see, according to your assumptions about ratings, the chess player who plays 18,250 games rated games per year, and wins them all, should be rated MUCH higher than Gary Kasparov who only plays 50 games a year and wins them all. Utter nonsense.

Blackice, do you finally see how invalid your assumptions about ratings are? I don't know any other way to explain it.

[blackice]

No I have a handle on it as per my example in the last thread. I have links to the uscf and several other chess ranking site and forums. I have read extensively about their rankings system. Don't assume so much I have also read extensively on the Tennis ranking system that is even worse. And once again CHESS IS NOT CTP bottom line. While both have "skills" involved one has a lot more "luck". You simply can not dissmiss that fact because you want to.

My examples you ignore you can not compare chess to CTP. They are completly different in so many ways it's not funny. You claim playing more games will get you higher in the ranking here. You assume that they win every game they are in they don't. The base of your points are based on assumptions pure and simple.

You can not base a "pure skill" ranking system in CTP unless you completely remove the variables as they did in the tourney. Is this what you want to do too to make the old system work? Make a new rule to remove the variables so we can simply play based on skills?

If not then the "pure skill" ranking system does not work for CTP. It does not account for "luck" pure and simple. The new system was used in GL for years tired true and it works.

Brunetti definitely has the best rating here, but when I looked at his past games he is 38-16-24 against players that are 1900+. This is impressive, but I then looked and noticed that he is 9-0 against craven who is rated 2002, but he is 5-10-1 and has never beaten anyone better than 1713 and is only rated highly because of his games against Brunetti. This would make Brunetti 29-16-24 against 1900+ players which is still good, but he is 574-19-38 total, which means he is 536-3-14 against players currently rated under 1900. I know he accepts all challenges, but this is quite a big difference. I have also found that his rating against the rest of the top 10 is 3-8-6, with him being 0-4-0 against carlosmart, 0-1-3 against mestrinho, 0-1-1 against cairo, 2-2-2 against dj222, and 1-0-0 against ardeshir. I would also like to see him play against more of the top ten and twenty players, but that may or may not be his fault. The idea of this post is not to dispute whether or not brunetti is good or not, because he is definitely much better than me and most other players here, but it is rather here to inform others of the facts and find out who people feel is the best player here.

There is your utter non-sense Quinn straight From gameknot chess site. This can not happen with the new system. Get that in your head please. You can not play weaker players, many of them and soar in the rankings.

Understand you wage points here lower ranked players add so little to the pot it simply is a nonsense scenerio you assume. You may win 10 of them and have the equivalent points gained from one well balanced high ranking game. NOT one player ranked right now fits your bill none. I do not see it happening.

I DO HOW EVER see people quiting games and playing the one they are winning in to become the next Klair with the old system!

Hey maybe Gary Kasparov plays but the old ranking system which would explain why he play so little and remains on the top spot. Either way I care not about chess in a CTP forum..

The new system awards for skill Quinn you do not win an entire game based on luck ok.

A higher ranked team must put up more points since a higher ranked team should have a better chance of winning. By solving the rankings in this manner it is not easy for a team to cheat the system. A team cannot play a weak schedule and expect a high rating because of a good record. Conversely, a team cannot play a tough schedule and expect their schedule to get them a high rating.

Many Differential rankings systems use this method. The new system the higher ranked puts put more points and obviously has more to lose. Why can you not see this new system is built upon the "Differential ranking system" I think the key lies in keeping risk and reward balanced. The new system does that the old does not.

http://www.geocities.com/TimesSquare/Hangar/5176/internet/ratings.htm

Quinn there is a good comparison of FIDE and the Elo rating system. You know what both of those are. We here fall into ELO and the same results happen (like Klair) here. If you want an exact FIDE system you need to apply the same methods. That would include taking all the "luck" factors out of the game. You really want to do that? Count me out I do not want to play a watered down version of CTP just to make a system work. And like most here I would prefer to pick the games I am in thank you. You will see it clearly states " More loss points than win points." Again you assume way too much about what I know and do not know. You can stop anytime you like. Now why should I be awarded more points over a stronger opponent if a third person took him out in a game? That is not meassuring squat about my ability to beat him is it? No this is not chess is it...

I also encourage people to go read this article because it clearly shows what the old system was like!
We here would compare to the ELO system and all it's pit falls which I have been making clear for what two friggin years...Klair factor all over again unless we increase the rules bah!

Just use a differential based system that does work for CTP was designed for CTP and let's move on.

[quinns]

Uggh....

[Solver]

Uggh....

Exactly my point of view.