 |
| |
|
|
Solver
|
|
Apolyton Duke Of Something
|
 |
Latvia, Riga
Sep 2000
time: 07:12
|
|
This is just an idea, but it seems interesting for me, after seeing something like this on some other sites and the MSN Gaming Zone. The idea is: all the played PBEMs are recorded by a person. Let's call him Ladder Admin.
So, for example a PBEM game has ended. At least 3 participiants of the game (for securtiy) send results to the Ladder Admin. I have created an evaluating system some time ago, but I know it's not ultimate and need improvement. Basically, this depends of: Map size, level, number of players and type of win. For example, email from the gamers might look like this:
Giant Deity with 6 players:
a - loss
b - loss
c - win (Alien Project)
d -loss
e - loss
f - loss
Well, I think you got the idea. So far this is only theoretical, but if you like this, I will try to make this real.
And so, with new points being added to / deducted from your score, we get the new ladder. Actually, I know that PBEMS take lots of time, but they do finish sometimes, I believe! In any case, I think it would be quite interesting to keep such a ladder, so that you can check it any time. Basically, I just wonder why isn't this done at this forum before.
------------------
Solver - http://www.aok.20m.com
|
|
| |
|
|
|
Stavros
|
|
Germany
Jun 2000
time: 05:12
|
|
I hate to be a party pooper, but just look at how many PBEM threads have ended in a winner. At best, people give up when the winner is clear... If I have missed something, I'd love someone to show me a thread where a PBEM has produced a winner.
But someone has a site up and running that has some some statistics of the PBEM games. I can't remember where/who has it. It had things like time between turns, and stuff like that. Can anyone remind me where that is?
[This message has been edited by Stavros (edited October 11, 2000).]
|
|
| |
|
|
|
deleted
|
|
formerly of Jack
Oct 2000
time: 05:12
|
|
Slamp won Modquick: he won it by conquering the World. It was on a small map and was a scenario started in the Modern Age.
Two of the Deathmatch Moderns finished and Im sure some of Lung's Deathmatches finished with complete extermination of all but the winner.
AFAIK as I know those are the only games which ended with a non-conceded win.
|
|
| |
|
|
|
quinns
|
|
Solver, I think this is a great idea!
Here's one way it might work.
Once the game starts and is agreed to by all to be a "rated" game, then the CTP PBEM Administrator, (I nominate Solver!), is notified by the "game starter" that a rated game has started (on such-and-such date and time). The courtesy turn around time would be 24 hours, from the time one turn is sent to the next player. The turn is delinquent if it goes past 72 hours. As soon as the turn becomes delinquent, the "game starter" would contact both the delinquent player and the CTP PBEM Administrator to warn of pending ejection (if it is the game starter that is delinquent, then some other player would notify the Administrator). If the turn around time exceeds 168 hours (one week), the delinquent player is automatically resigned from the game and that player's nation is "permanently" replaced with a Computer player (AI).
The last "human" player remaining wins the game. Order of finish is in order of elimination (either by conquest or resignation). If more than one human player remains when the game ends, then those remaining human players are ranked according to the civilization score (regardless of what the AI did). For example, in a 5 player game, the first player ejected comes in 5th place, the second player ejected comes in 4th place, etc.
This way there will "always" be a finishing order to the game, no matter how short or long.
I own a site called "World Tennis Ratings" http://www.worldtennisratings.com . I am very familiar with ranking and rating systems. I could fairly easily modify that tennis rating system to adapt it to work with multiplayer games. We could start everyone off with a rating of 15.000 to begin with, then have the Administrator initially rate new players as they join the CTP PBEM Rating system. Ladders are then easy, as they are just sorted by descending rating.
If there is any interest in this, I would be happy to elaborate on how the rating system would work.
Steve Quinn -- steveq@gte.net
[This message has been edited by quinns (edited October 23, 2000).]
|
|
| |
|
|
|
Solver
|
|
Apolyton Duke Of Something
|
|
Latvia, Riga
Sep 2000
time: 07:12
|
|
Quinns, thanks. I can be CTP PBEM Admin, as I have some admin experience thanks to Age of Empires II. So, perhaps we should make the game we just started (how about some?, I mean) a first rated game. If you want, I will post my ideas about ranking system, but these are not very good, so I want your ideas. I would be happy to see more support, as your idea about how to make each game finish is great. My idea is great, too 
------------------
Solver - http://www.aok.20m.com
|
|
| |
|
|
|
Solver
|
|
Apolyton Duke Of Something
|
|
Latvia, Riga
Sep 2000
time: 07:12
|
|
Aha, I have fond something. Again, I invented it some time ago and I don't say this is what we should use, just a s an example. It's in HTML, so I will now transfer it here.
Then, when you win a game you get:
100 pts if you wan by conquest,
150 if you won by having highest score and 200 points if you won by
creating an alien. If map size was middle, you get extra 50 points,
if it was big, you get 100 and if it was largest, then you get 150
pts. If difficulty is Warlord, you get 25 pts, 50 for Prince, 100 for
King, 150 for Emperor and 250 for Deity.
If there were 3 players in the
game, you don't get any extra points, but you get 10 pts for 4
players, 20 pts for 5 players, 30 pts for 6 players, 40 pts for 7
players and for 8 players you get 50 points.
Example: If you won by Conquest
on Prince difficuly, with 6 players in the game on Middle size map,
you get 100 + 50 + 30 + 50 = 230 points.
Example: If you win by creating
an alien at Deity level with 3 players on the biggest map, you earn
200 + 250 + 0 + 150 = 600 points. Makes difference, doesn't it?
Also, if you lose, points are deducted.
If you were conquered, 200 points
are reduced, if your enemy had the highet score win, 150 points are
reduced and if he created an alien, 100 points are deducted. For big
map, extra 100 points are reuduced and for largest map 150 points are
deducted. For Warlord you lose 200 points, 150 for Prince, 100 for
King, 50 for Emperor. For Deity no extra points are taken. If there
were 3 players, you lose 50 points, 40 points for 4 players, 30
points for 5 players, 20 points for 6 players, 10 points 7 players,
and you lose no extra points for 8 players.
Example: If you lost, because
your enemy created an alien, on King level, largest map, with 4
players in the game, you lose 100 + 100 + 150 + 40 = 390 points are
deducted from your rating.
First of all, sorry for the spelling mistakes (it must be full of them), but take a look at this.
Quinns, haven't you by chance played games on MSN Gaming Zone?
------------------
Solver - http://www.aok.20m.com
[This message has been edited by Solver (edited October 23, 2000).]
|
|
| |
|
|
|
quinns
|
|
Here is my take on the CTP PBEM Ratings.
Make them as simple as possible. I would strongly recommend using rating adjustments, similar to World Tennis Ratings (as this is tried and true, and in the process of possibly replacing the existing professional system). World Tennis Ratings is loosely based on the International Chess Rating system, but much simpler and cleaner.
Modified Proposal:
Player CTP PBEM
Level CTP PBEM Rating Ratings Color
E 0 through 9.999 RED
D 10 through 13.999 ORANGE
C 14 through 16.999 YELLOW
B 17 through 19.999 GREEN
A 20 through 22.999 BLUE
OPEN I 23 through 25.999 BRONZE
OPEN II 26 through 29.999 SILVER
OPEN III 30 and above GOLD
If we use this structure, I already own the programs that I could modify slightly to work with PBEM games.
We could give the following initial ratings to players that have honestly defeated the AI at different levels:
Defeated Initial
CTP AI Rating
Chieftain 14.000
Warlord 16.000
Prince 17.000
King 18.000
Emperor 19.000
Deity 21.000
The ratings would change using the formula: Probability of Lower Rated Player Defeating Higher Rated Player =
Prob = 1 /(2 + (player1 - player2)^2).
If the higher rated player defeats the lower rated player, then the Change in Rating = Prob./2
If the lower rated player defeats the higher rated player, then the Change in Rating = (1 - Prob.)/2.
see www.worldtennisratings.com/wtr_info_formulas.html for a better explanation of the formulas.
Example:
Before Game After Game
Player Rating Finish Order
Joseph 17.000 2
Michael 20.000 1
Samuel 22.000 3
The above formula is applied each time a player is defeated or resigns, using the CURRENT ratings.
1st Pass -----
Samuel was ejected first, so the first pass calculation goes as follows:
Joseph Defeated Samuel:
Prob. = 1/(2 +(22 - 17)^2) = 0.037
Change = (1 - 0.037)/2 = 0.482
Joseph's new rating is 17.000 + 0.482 = 17.482
Samuel's new rating is 22.000 - 0.482 = 21.518
Michael Defeated Samuel:
Prob. = 1/(2 + (21.518 - 20.000)^2) = 0.232
Change = (1 - 0.232)/2 = 0.384
Michael's new rating is 20.000 + 0.384 = 20.384
Samuel's new rating is 21.518 - 0.384 = 21.134
2nd Pass ---
Michael Defeated Josesph:
Prob. = 1/(2 +(20.384 - 17.482)^2) = 0.114
Change = (0.114)/2 = 0.057
Michael's new rating is 20.384 + 0.057 = 20.441
Joseph's new rating is 17.482 - 0.057 = 17.425
---- End of Calculation
After Game
Player Ratings
Joseph 17.425
Michael 20.441
Samuel 21.134
--- End Example
The computer program will do all of the above work each time a player is ejected from a rated game. I could help you with this part of the administration.
I would also recommend using these same exact formulas regardless of what scenario is played. Because it prevents tampering and abuse. As you mentioned earlier, if we give more points for complexity and level, a group of players could initiate a game with the highest possible point levels, then just have everyone quit after one turn. All the players would reap the benefits of the higher point game without really playing. With the above proposed system, players could play one turn or one-thousand turns, and the system would still calculate fairly. The only real variables are the players' ratings and players' finishing order. (At least one turn must be played in order to be considered a valid game).
So much for my long winded explanation.
Let me know what you think 
[This message has been edited by quinns (edited January 19, 2001).]
|
|
| |
|
|
|
Solver
|
|
Apolyton Duke Of Something
|
|
Latvia, Riga
Sep 2000
time: 07:12
|
|
Amazing, Quinns! Are you a programmer? If so, and the program is your, please send the source code to me, or if it is in Visual Basic, send all the project to me, as I can understand it. Well, I guess I'm not so good at understanding all the formulas, however, this seems very good. Just a question: with how many rating points should everyone start, so that this seems fair and good. And, how do americans type four thousand: 4.000 or 4000? I'd like to know this, as in Europe putting . means it's part, i.e. 2.5 means 2 1/2 , so let me now. Well, I guess our game How about some? should be the first rated, I really want so. Ideas? I will wait for email with the code and reply here.
------------------
Solver - http://www.aok.20m.com
|
|
| |
|
|
|
quinns
|
|
Hello Solver, I'm glad you approve! To answer your questions:
Question: "Are you a programmer?"
Yes. I program/analyze for a electronics manufacturer in California. The languages I used for the World Tennis Ratings site are a combination of PickBasic, HTML, JavaScript, and Perl. I know only a very little Visual Basic. (At work we use PL/SQL only, now  )
Question: "With how many rating points should everyone start, so that this seems fair and good?"
As stated above, I think that, at first, everyone should honestly state what level they believe they play at, Chieftain, Warlord, etc. Then, you (The Administrator), will make the official determination of what Initial Rating they should have. (See above rating guide.)
Question : "How do Americans type four thousand: 4.000 or 4000?"
We type it just as the Europeans do. In America, 4000 is written as 4,000 or 4000. 4.000 means four decimal-point zero zero zero (or the number four exactly). I think the British use a comma to represent a decimal point, but not the Americans. In reference to the rating point range, it should run from about 5 to 40. I can't imagine anyone attaining a rating higher than 40.
Comment: "Our game, "How about some?" should be the first rated game, I really want so. Ideas?"
I agree. Let's see with the other players think. If they all agree, then let it be so
[This message has been edited by quinns (edited October 24, 2000).]
|
|
| |
|
|
|
Solver
|
|
Apolyton Duke Of Something
|
|
Latvia, Riga
Sep 2000
time: 07:12
|
|
Unlucky. Well, I guess I can understand codes for all the Basic programming languages,as they all are similar to VisualBasic, however, I can't understand Perl. Actually, I've never seen it, so perhaps I will understand something.
Well, Quinns, I want you to post in our How about some? game thread and ask the others would they agree, and direct them to the thread. I would do this myself, but I'm in a great hurry now.
I will wait for further posts by you.
------------------
Solver - http://www.aok.20m.com
|
|
| |
|
|
|
quinns
|
|
Okay Administrator Solver! Will do. Forums posted.
|
|
| |
|
| |
|
|
|
quinns
|
|
Re: "How often to post?" -- How does once a week, on Sundays, at 2300 GMT sound? In the beginning, I could just manually calculate it, but if this gets popular, I'll have to create a new program or maybe you could do it. (If you want to see an example of the rating change calculation in JavaScript, use your Browser to go to -- http://www.worldtennisratings.com/wtr_info_calcrtg.html and click "View -- Page Source". You'll see the JavaScript code for calculating rating changes and odds for two given players' ratings.)
Yes, a new topic under this forum sounds good.
If Arthur and St Jon agree, we need to establish initial ratings. How did my earlier idea sound? That is, regarding initially rating a newly rated player dependant upon what level of CTP AI the player has honestly EVER defeated, (in a single human player game). (See above thread).
For example, the highest level that I've ever honestly defeated the CTP AI was EMPEROR that would give me an initial rating of 19.000 (or 19 exactly). This would, of course, be subject to approval from (you) the Administrator of CTP PBEM Ratings.
[This message has been edited by quinns (edited October 25, 2000).]
|
|
| |
|
|
|
quinns
|
|
Regarding Cheating. There's no way to stop a cheater other than to just blackball them if you think they cheat. (It seems to work in other games and sports.) I don't even think that this is a good method, really, because you could be wrong. Sometimes it might look like someone cheated because they are way ahead, but they didn't cheat.
It's probably pretty easy to cheat in PBEM if somebody wanted to do so. And someone could just lie about what level they defeated the CTP AI to get a high initial rating. (Actually, the rating system will eventually correct this artificially high rating when the player actually competes against real players.)
It's a good question...
They have the same problem with online chess. I belong to the ICC, an international on-line chess club. I have been accused of cheating when I was not. The ICC has been doing this for a while, so they actually can tell whether somebody used a computer to assist them while playing (i.e. cheat). I was accused and acquitted by the ICC.
The only way to do it is to trust everyone that participates, and banish anyone that has "solid proof" against them of cheating. Until that time, everyone will be trusted. I don't see any other way.
Yes, if you agree with this method of initially rating players, then we need to get St Jon's, Birdman's, and Arthur's best solo finishes against the CTP AI.
|
|
| |
|
|
|
quinns
|
|
Okay Solver. That's it then, we won't change it after this. I'll edit the previous posts to avoid confusion regarding starting ratings.
**Official**
Starting CTP PBEM Rating Table
(Based Upon Best AI Ever Defeated in Call To Power (as a single human player)):
Chieftain 14.000
Warlord 16.000
Prince 17.000
King 18.000
Emperor 19.000
Deity 21.000
[This message has been edited by quinns (edited October 26, 2000).]
|
|
| |
|
|
|
Solver
|
|
Apolyton Duke Of Something
|
|
Latvia, Riga
Sep 2000
time: 07:12
|
|
Yep, agreed. We are at the moment waiting for people to give their results. So far we have only 2 of them, but 3 more are coming just in moment:
The BirdMan 21.000
Quinns 19.000
Solver 18.000
Soon to come: StJon, Arthur.
Quinns, there's another game pending - I mean "Want to start CTP PBEM but no Idea about how to work it". This game will be rated as well, so we just need to wait for the scenario. I think the game is about to live, as there already are three players: Arhcr-Druid, Quinns, Arthur and Solver, perhaps fifth one will appear.
------------------
Solver - http://www.aok.20m.com
[This message has been edited by Solver (edited October 26, 2000).]
|
|
| |
|
|
|
Stavros
|
|
Germany
Jun 2000
time: 05:12
|
|
I'm sorry, I'm still sceptical 
First I'm going to write a disclaimer so you don't think I'm just a whinging Aussie... then I'll let loose
I like your idea of a ranking system. It would be great to be able to see some sort of relative comparison between all the colourful characters that we have in this Forum. I'm not trying to discourage or humble your efforts, but I have a few things that are causing me concern:
1. The length of time it takes to end a PBEM game. (I have got it right haven't I? This is only for PBEM?)
You want to post the results once a week? How often does a game finish? I think once every 3 months would still show a pretty static ranking....
2. The games with more than 3 or 4 players (most of them) are inherently biased to the first players to play their turns. They get the rivers, the good land, and generally speaking a better starting position. The exception to this are the games that have been 100% created for a set number of players, and where the the creator doesn't play. I don't think there are any of these at the moment, are there? Don't get me wrong, I still enjoy playing in these biased games, but only to be part of the fun. No matter how good a player I am, I am not going to win a game where Player #1 has a river 20-30 squares long, and Player #8 has a 2 square swamp river, surrounded by desert. (This is where I am in Sows "Do Some PBEM" btw.)
3. For a ranking system (ala Tennis) to work, doesn't there have to be a large turn around in results to flatten out anamolies. I'm no expert (only a couple of years of Uni. Statistics), but having looked at the formula I see that in this case the reverse could indeed be true. It will take a "settler" about 5-8 games of whipping everyone else's but to get up to the rank of someone starting on Diety (presuming they don't win a couple as well). By the time he catches up to these players, I think that we all may have started CTP2  My suggestion here would be to start everyone on the same score. If the ranking system works the way it should, it will sort itself out, shouldn't it?
4. Just a thought really, on how we might be able to get more intermediate results. For every "ranking" game, every (x) turns, we could perhaps look at the relative rankings of the players in the Powerchart. I know it's not strictly speaking an acurate indicator, but it's the only one we've got Using these relative positions, we could fudge together some kind of algorithm to get a score (I'm sure the one you've suggested would work here as well) This doesn't solve point #2 (inbalance in starting position), but deals with 1, and 3 quite nicely.
Just my comments, and (I really mean it) keep up the good work...
|
|
| |
|
|
|
Aaginor
|
|
Germany
Apr 2000
time: 05:12
|
|
Hi Folks.
I also see some problems with the ranking system.
First is the time to complete a game. Maybe we should think about stavros suggestion to count the score after each x turns.
Second is (as Stavros also mentioned) that it depends on the starting position, how you are doing in the game. You can be a freaking crack ... but this doesn't change the swamps and deserts ...
Third is: What about the subs? If you join a game, when it is already runnning, and you are taking a weak civ, you will be punished for the faults of the former player.
|
|
| |
|
|
|
quinns
|
|
Skepticism makes for a stronger system.
Let me try to address some of these issues and see if we all might find some common ground to agree on a single system.
--------------------
1. Uneven Starting Position --
The only way to avoid this, is as you mention, to have a completely even map, (known to all), that is symmetrical (3 to 8 ways). The map would look more like a Parchesi board (kaleidoscope?). That might be interesting and would avoid anybody from cheating by looking at the revealed map during play. However, I'm for leaving the map random, and letting the players' skills try to handle the hand that's dealt them. It would be part of the Administrator's duties to try to alter the starting order for players to make it fair, (e.g. if a player started last the last game they should start closer to first the next game.)
2. The Ratings Don't Change Often Enough --
I think they will. As mentioned before, there will be no substitutions. If someone doesn't respond (within the one week deadline) then that player is resigned from the game and permanently replaced with an AI player. All of the players' ratings in that game would be recalculated AT THAT TIME (based on everyone defeating the resigned player).
3. Starting Rating Should Be The Same For All --
Maybe this is true. However, in your example, the "Settler" should have stated a higher starting rating (like King or Emperor). Even in the worst case, imagine a game where a player rated 14 plays against 4 other players in one game that are each rated 20. If that 14 player wins, then his rating would be increased to 16 after only one game (and the other players would drop to about 19). After winning two games, the 16 rated player would be at 18.
One more point. I've found that if everyone starts at the same level, it doesn't give a good range of the rating numbers. Everyone seems to hover around that same starting rating. The better players will have higher ratings, but not much higher than the worser players.
-----------------------
Again, I hope we can all find some common ground on this. Awaiting your replies...
Respectfully,
quinns
|
|
| |
|
|
|
quinns
|
|
Stavros -- "The switching positions can be tricky, because it is usually arranged by time-zone to improve game speed."
Quinns -- Yes, agreed. But I could see just starting in the timezone of the "most-last-started" player and moving westward in following games. For example, I started last in our game, R1. Next game, I would start first (United States, Ca) and just move westward in the time-zones -- (Australia, Russia, Germany, England, for example).
Stavros -- "...But the uneven start is a general problem, and probably too large a task for us to tackle in this thread..."
Okay. Does someone have any symmetrical maps created out there? I think this is the best approach to really fair starts, (if that's what we want to do). The map terrain would have to wrap perfectly symmetrical around an Earth shaped world base upon how many players were in the game.
Stavros -- "...So in the perfect game (which we all strive to create where nobody drops out (unless beaten) will still suffer the same problem [of infrequent rating changes]..."
Quinns -- My take is that Rated games should be played a bit differently. I would encourage more resignations by players when the power graph gets too obvious, rather than playing to the bitter end, (there's always another game). And if people are playing 5 to 6 rated games at a time, there may be 2 to 3 games with dropouts each week (not necessarily finishes) which would cause the ratings to change.
Stavros -- "The second issue that I see is the replacing of the player by AI. Many players prefer substitutes due to the fact that the AI is so weak. The substitute could perhaps continue to play, but not be included in the ranking system."
Quinns -- I would have to disagree with this for almost the same reason as the previous issue. We really WANT the "rated" games to end so that we can cause dynamic rating changes. Let the AI be weak and get slaughtered by whoever wants to risk it. I see the "unrated" games having more role-playing and tapping into the more exotic side of CTP, and those games will still go on. I still vote for "permanent" replacement of resigned players by the AI. (It sure simplifies things.)
Stavros -- "As an aside, what is your opinion of updating the ladder every, say 20 turns (this is after generally 1-2 months game time)? I know that the powergraph is no true indication of the players prowess, but in the end, they come out on top... This would even allow permenant substitutes to continue in the ranking system."
Quinns -- The formula, as it is, won't work this way. It needs to see clear winners and losers along with each of their corresponding ratings. That's it. How did you see doing this? Does everyone have a civ score after 20 turns? Even so, if we go this route we would have to come up with a different method of rating, altogether. I couldn't modify the system (proposed above) to work with "civ points". I'm not against another system. I'm just throwing this system out there to see what comes back.
Best regards to all!
[This message has been edited by quinns (edited November 01, 2000).]
|
|
| |
|
| |
All times are GMT. The time now is 05:12.
Apolyton Time is 00:12. |
top of page
|
| archivepost |
|
Forum Rules:
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts
|
HTML code is ON
vB code is ON
Smilies are ON
[IMG] code is ON
|
|
|
|
|
|
11-10-2000 09:41
|