Ctp Pbem Rankings 12th April 2002

[blackice]

How is it "flawed" it would help to understand better if you gave examples. Klair BTW was just one example as you know there were more but that is old ground now.

In this system now has a clear winner and a loser and everything in between is graduated. The winner takes the lions share of the pot as it should be. The ones in the middle lose or gain a bit but are not wiped out in the rankings. The loser is hit but again not wiped out in the rankings.

Seems fair to most and has the rankings moving again. Lung can not sit on top without losing his spot if he does not play more games and stronger opponents. That is primarly the only spot that feels that effect. The rest contiunue to play as per usual and the postions move quite frequently upwards and some slightly backwards. But no great dives.

[Keygen]

Well, I think that the "Klair" factor could easily be eliminated by setting a minimum limit on how many rated games someone should be in to be allowed to be rated. For example he/she should be in at list 3 or more rated games before can make it to the ladder.

Now didn't I say something about starting a new thread and posting the old and new rules there?
Perhaps I should do that myself too?

[Solver]

Uh ah, scary Keygen, you can do it . I just have to get the guts to write all the rules.

Also for the "Klair factor". A penalty for not playing a long time could also be eliminating this. And 3 game limit is fine by me, too. Not more, though.

[blackice]

Ok once again you eliminate people that want to play casual. I see no need to create this stigma we are a gaming forum not a hard core gaming forum. Lots of people here have two games and no time for more.

I don't get it you have a ranking system that works yet you want to go out of your way to make it hard on people. Or eliminate people to make an old system work. Or make rules to make it work dancing maddly backwards I say. Klair as posted is not the only reason the system was flawed so you need more...

I have not heard one good reasons so far to go backwards anyone else?

So again the question is how can you justify changing a system that works. I have challenged you all for reasons yet I get nothing. That speaks volumes....

Solver has spoke he said for all intents "because I can" that is a good reason is it not? Solver also told me to use Gary's system...Politics obviously.

But go ahead people make new rules to eliminate potential players to make an old flawed system that does not work, work

Stick your neck out start a poll and give the reason's for a change if you dare. Chances are it not only will not happen but the people who like this system will be over ruled.

Lots of politics no common sense....

[ricketyclik]

Here's my two cents worth...

I see no problem with the present system, except the obvious one which is that it is not being applied (updated)

Another ladder I play on has the system where if the winner is ranked below the loser, the winner moves half of the places toward the loser's rank. Loser stays stationary. Ranking is reduced as other players leapfrog past you.

If you're in a 5 player game and on top, you get 4 wins, if you're in say 4th place, you get 3 losses and 1 win. The system has the effect that if you are inactive you slip quickly, unless you're near or at the top, in which case the loss of rank due to inactivity is necessary, although not applied in the ladder I play on, which is frustrating.

Again I say: the present system is fine, it just needs to be applied

Edit: By the way, hello from the sunny Philippines

[blackice]

Similar to the old system Rick. If you are in thrird place in each game you play you get two loses. Fourth place three some incentive. If you are first place in one game but third in the rest you nose dive down the ranks big time.

This system the third ranked my move up or down depending on the pot. They may also stay as they are.

So in the example above you may gain rankings depending on the pot not just nose dive to the bottom which in the old system was assured.
Klair was just one minor flaw...

This system is far more fair as far as movement and interest goes.

As far as stagnent players go look at the stats the top players that stopped playing are falling fast. No need for a penalty there they are gone.

I am still looking for other than politics to change the current system.

Add it up people all and all a more fair system and no need for inactive penalties. This system adds the incentive needed not the penalities which is negative to begin with.

Thanks for your input Rick come on people you do have a say I think?

None the less post your thoughts

[Keygen]

I am not sure whether you are refering to Solver or me or both on your above comments Pete but I feel like I should properly responde to you anyway.

I have persistently asked for both the older Quinns/Solver and newer Gavrushka/Blackice full rules. If I remember correctly the latter had never been posted here but if I am wrong correct me. I didn't have the time to do a search on my own yet so if Pete could do that for the latter and Solver for the first since you possibly both will know where to look at first in the contrary to me I would greatly appreciate it.

Now Pete if you can straitly answer to me how your system or Solver would handle the following situation I would be more than glad to hear about it.

Someone plays in 8-10 games. He's a good player but not excellent. He's rated only in three of his game. He took over these games as a replacement in 90-120 turn when his predecessor was an excellent player and had a way ahead lead. In his rest games he is in the middle or bottom of the power graph. Another excellent player is playing in 5 games. He is rated in 3 of them. He leads in the first two but he's second position in the other with a significant distance from the first one. The same for his rest unrated games. Now supposedly all 6 rated games from both player is more or less the same how the hell will ever the excellent player going to beat the good player? Note that a good player can keep a huge lead from his predecessor in the above advanced stage that took over as a replacement perhaps up to the end of the game. Only other factors will give the excellent player chances to pass ahead if the good player get attacked by combined forces and be beaten (the most likely) and the excellent won't get attacked.

If you can solve this problem I will withdraw all my suggestions and pronounce your system excellent.

If you beleive that a system can be perfect from it's first release or second release and no further involve can not be made then I guess the 99% of the computer industry are dump!

Now I only intented to do some suggestions nothing more. Before you turn them down you should first pay more attention to them. For I care less for being rated and that can be easily seen from my lack of posting on the ratings for the past 9 months or so. Never was particularly interested on them and only agreed to get rated to help out. Mostly interested on helping with the rules but every time I tried to found myself under fire in any of my suggestion, not to say that nobody's else suggestions also ever accepted. Not sure what the fuzz with the ratings and everybody involved with the only expetion Solver (for the first time) are so hard to convise them that their work doesn't work excellent and could use a couple of suggestions to make them better.

Now do not take my post as offence, I am not here to offend anyone or fight with anyone, I find it completely useless but had point out a few things.

Now can you please keep me out of the ratings?

Thanks

[blackice]

It think there was a rule for that scenario somewhere. None the less the rules were not changed just the rankings system. In the old system both players would drop. We saw this all the time ask ber, faded, Lung or me. It did not go as far or in a fair way of promoting many games. As Klairs example that was the best way to stay on top.

In the new system they may or may not change a position. It depends on the rankings of the other players and how much is wagered in the pot. It does promote playing one or many games. Many games is obviously going to get you to the top faster but is that not how it should be? Mind you in this new system playing one game and winning it every ranking will keep you up there but not top spot. Also playing many games does not assure top spot either. Your ability does and your opponents strength. A strong player playing weaker ones may win but the share of the pot may be weak. Sooner or later this good player has to play stronger players to get any movement going. This works well here because generally games are made up of stronger and weaker players.

I think we have to be very clear here the "rules" and the "ranking system" are two different things. I agree with you Keygen some "rules" need to be updated and or spelled out better. But stress again the "rules" were not changed just the "ranking system"

I hope that clears it up seems we have been mixing the two

[Kralj Matjaz]

Only true difference between systems is that first one (Quins/Solver) do not reward playing a lot (which is correct) and second one (Gawrushka/Blackice) promote playing a lot through inflation points. In both systems you must keep wining, but in second system is enough to play a lot against average/week players and be on top while in first system you must keep wining against the best.
Both systems can be improved. First one with not allowing one "lucky" game decide (only real flaw as I see it).
In both systems old points should decay through time (year?).

[jpww]

Originally posted by Keygen
Well, I think that the "Klair" factor could easily be eliminated by setting a minimum limit on how many rated games someone should be in to be allowed to be rated. For example he/she should be in at list 3 or more rated games before can make it to the ladder.

I agree with this...

Originally posted by Kralj
Only true difference between systems is that first one (Quins/Solver) do not reward playing a lot (which is correct) and second one (Gawrushka/Blackice) promote playing a lot through inflation points. In both systems you must keep wining, but in second system is enough to play a lot against average/week players and be on top while in first system you must keep wining against the best.
Both systems can be improved. First one with not allowing one "lucky" game decide (only real flaw as I see it).

And this.

All in all, I favor Quinn's system in conjunction with setting a min. # of games to join the ladder. Additionally, I like Birdman's idea of rating each player in every game for calculating ratings more accurately

[blackice]

Kralj Matjaz-Lung can not play weak players and stay on top no one can. He was making no points in the game he was in. In fact he was getting negative points for playing weaker players. This system too promotes playing strong players but is balanced by the fact most games have a mix of players. This system allows for that the old system did too but the more games you played the more likely you dropped in the rankings. No one is that good or that lucky to have good land in every game. In the old system one loss was as good as three wins. That made no sense at all. You play three games win one lose two you drop in the rankings big time. Many positions usually that is why no one could catch up to Klair. It made no difference how many games you were in your losses were three times that your wins.

With the new system that simply does not happen. Yes you take loses but it does not drop you from 7th place to 15th in one ranking like the old one. You may drop a spot or even two max with the new system or not move at all depending on the ranking of the other players. If you play higher ranked and lose you go down. Average players it may not make a difference to you but the average ones more than likely will gain points for a second or third ranking. A win win scenario for all players...

As far as decay goes look at Klairs and other retired players standing it will take a few months and they will be all but gone with this system. The old one more than likely they would be still there.

jpww-what you are saying is in order to join this league you must play X games. Why would you want to eliminate half the players here and a whole slew of new people? Makes no sense at all the numbers here will more than likely shrink.

As far as Birdmans idea what you are saying is you must play rated games or go somewhere else to play? What is this the elitist's club?
If someone wants to join a game play with the people here why should we put rules on that? No you can't play go away because they do not care for rankings systems? Or they get bad land and quit because they do not want to take losses in every game. We need to promote play here even if they are not ranked they may decide to keep the bad land game going to learn from it. Or decide in the future when they have more time to play more games. With the old system they were digging dirt. With the new one they may not be top spot but chances are they may stay the same or improve in the rankings be it slow it is a positive not a negative. They are still here having fun though and you want to stop that? Maybe you could explain why you would do that? People come here for fun not more rules and not to be told they can not play because they do not want to be ranked... To much time has be put into making people play by a ranking system. Why no just let them play a game ranked or not?

JP do you even know how quinns system was calculated? Kralj you? My guess is niether of you do not many people did. If that is the case how are you making an informed decision?

Better yet if you do explain to the rest of us...or how about quinn or solver taking the time to spell out here with examples of the two systems so the newer people can understand? How about it put them side by side and compare them. Lets do it so we can all see the difference we are talking about here. Enough with politics lets compare them side by side...

It is easy here it is the old system:

1st place in for player game gets three wins
2nd place one loss two wins
3rd place 2 loses one win
4th place three losses.

2nd through 4th will drop in rankings period up to 7 positions or more. For one game ranking on the 9th turn.
Unreal and unfair IMHO.

new system:

Pot
1st place may or may not gain depending on the players
2nd place same
3rd place same
4th place same

In other words no one really losses and rarely anyone will drop more than a spot for one game ranked on the 9th turn. The average player even in 2-3rd place players will come out with a few more points than lost or even. Even with the losses you do not drop several places in the rankings for 9 turns of gaming. It would take many many such turns to drop like the old system.
More fair and promotes more games on it's own without rules...

Vote new It's better!!

[Kralj Matjaz]

Originally posted by blackice
It is easy here it is the old system:

1st place in for player game gets three wins
2nd place one loss two wins
3rd place 2 loses one win
4th place three losses.

2nd through 4th will drop in rankings period up to 7 positions or more. For one game ranking on the 9th turn.
Unreal and unfair IMHO.

Considering old system it's not true at all! (well maybe in some very particular case).
First of all you play against each opponent separately. It's important not to be behind someone with much lower ranking or you'll really dive. And that is absolutely fair because he's not so bad if he beat you and most important difference between two will get smaller and in some turns both will end in the deserved place. Through time everybody will find his place (if it's on bottom instead on top too bad for overestimate player) on the ladder and bigger changes will occur only when new player will find his place.

Yes in the beginning there will be a lot of jumping up and down, but more games you'll play more accurately you'll find yourselves on the right spot!

Minimum number of plays (let say three) is OK here. There is no drawback. If not rated in X plays you are not on the ladder. Plain and simple.
If you don't wish to be on the ladder, that's OK. You can still play as usual.

My vote is of course with old system

[blackice]

Considering old system it's not true at all!

Kralj right ok show me i am wrong come on post it!

Take the old system and do the math for us do it I dare you.

I bet ten bucks US you have no idea how to calculate the old system make it 50...

now show me it is not that way using current stats...Politics and BS.

Yes in the beginning there will be a lot of jumping up and down, but more games you'll play more accurately you'll find yourselves on the right spot!

Right and klair was what a lot of games what ever!

The new system you play against each opponent separately too whats your point?

It's important not to be behind someone with much lower ranking or you'll really dive.

Ok you have a bad start you are behind DIVE BABY DIVE.
Oh and don't forget you must be rated, unreal...

In the new system you will take a hit too but not 7 ranking positions for one 9 turn game pathetic logic and a bad system. PERIOD.

You will take a hit and maybe drop 2 or more positions for one 9 turn game think about what was just said for one minute. You have the required? 3 games minimum you have bad land start in all games each friggin 9 turns of each friggin game you drop several ranking spots...
Oh ya baby bring it on...rediculous.

In the new system with the same scenario you will drop but it will take mega 9 turn losses to produce the same effect. Hey what does a game last for 150 turns 200+
That is 22 times say 3 losses in a four player game.

Wow the old system would have you in China. The new system ya you'll drop near the bottom but it would take it all to get you there. The old system you would be one quarter of the way to the 200 mark and they would give you concessions because you could not go lower without showing a minus on your score. Do the math POST THE SYSTEM show us you know what you are taking about,

Right! now I say you have NO CLUE...

Minimum number of plays (let say three) is OK here. There is no drawback. If not rated in X plays you are not on the ladder. Plain and simple.
If you don't wish to be on the ladder, that's OK. You can still play as usual.

heh brown shirts comes to mind. Why the heck not let someone play one game rated the new system can handle it. Oh ya right you need to make rules to exclude people to make the old system work...Unreal...

My vote is of course with new system most voting for the old system do not even know how it works
politics

Post it people show the reason you are making up rules to excule players to make an old system work/ Tell them there is no reason to do it when the new system needs no such rules to work. Explain to them the politics and why they can not be part of the rankings because of it...
Pathetic really...

[TheBirdMan]

I bet ten bucks US you have no idea how to calculate the old system make it 50...

Luckily you didn't "challenge" me Blackice. I made my own copy in a spreadsheet about a year ago on how to calculate the results - and it works. I got the same figures as Quinns did (actually Quinns did a fine job of documentation how it should be done).

One thing however, I did disagree with was where he took the input for each "match".

Something like this (3 player game):

Player A is no 1 and had a points.
Player B is no 2 and had b points.
Player C is no 3 and had c points.

Match "1" A vs B used player As a points and resulted in d points to be added to player A and deducted from player Bs b points, resulting in e Points to player A (a+d) and f points to Player B (b-d).

Match "2" A vs C used player As e points and resulted in g points to added to player A and deducted from player Cs c points, resulting in h Points to player A (e+g) and i points to Player C (c-g).

Match "3" B vs C used player Bs f points and resulted in j points to added to player B and deducted from player Cs h points, resulting in k Points to player B (f+j) and l points to Player C (i-j).

I wanted it this way:

Match "1" A vs B used player As a points and resulted in d points to be added to player A and deducted from player Bs b points, resulting in a total of e Points to player A (a+d) and f points to Player B (b-d).

Match "2" A vs C used player As a points and resulted in g points to added to player A and deducted from player Cs c points, resulting in a total of h Points to player A (a+d+g to be published) and i points to Player C (c-g).

Match "3" B vs C used player Bs b points and resulted in j points to added to player B and deducted from player Cs h points, resulting in a total of k Points to player B (b-d+j to be published) and l points to Player C (c-g-j to be published).


Edit: Those letters...... It's hard to explain without real numbers .

[Solver]

A short notice from me: I actually preferred the Quinns/Solver formulas because those are well tested by the Chess organizations. Gavrushka/Blackice formulas were good, too, but maybe a bit too simple.

The former formulas do generally provide a fair rating, and if we put a minimal cap on rated players, it will also eliminate the "One lucky game factor". Yes, we might get off a few casual players, but that's the price...