SMAC PBEM Ratings (01-Mar-01)

[Solver]

Alpha Centauri / Alien Crossfire Play-by-email Ratings

POS....NAME............ RATING

001....BIG_CANUK...... 21.881
002....MARK13......... 21.308
003....STUNTMAN19..... 21.167
004....SOLVER......... 21.133
005....REDFRED........ 21.000
006....DILTHIUMDAD.... 21.000
007....KEYGEN......... 20.875
008....TAU_CETI....... 20.228
009....REGGY.......... 20.160
010....MISOTU......... 19.842
011....GOOGLIE........ 19.145
012....BLACK_SUNRISE.. 18.542

Rated games in progess:

Keygen's Classic (2140):
1. Stuntman
2. Keygen
3. Nanuk
4. Reggy
5. Black Sunrise

CMN: MISOTU


Keygen's Alien Crossfire (2158):
1. Keygen
2. Reggy
3. Stuntman19
4. Black Sunrise

CMN: MISOTU

Keygen's Faction Strife (2101):
1. Keygen
2. Solver
3. Tau Ceti
4. Googlie
5. Big Canuk
6. Misotu
7. Mark13

CMN: Aredhran

> >[/i]

style='font-size:14.0pt'>> >[/i]

Frequently Asked Questions>>[/i]

> >[/i]

> >

Q: How do I become
rated?>>[/i]

A: It’s easy.
Just post, to the current ratings thread, the best SMAC AI level you have ever
defeated. Thus, you will get an initial rating depending on it, here the values
are: Citizen = 14, Specialist = 16, Talent = 17, Librarian = 18, Thinker = 19,
Transcend = 21.>>[/b]

> >[/b]

Q: How do we make a
game rated?>>[/i]

A: First, clarify
whether the players in the game want to be rated. At least three players must
agree to be rated, otherwise we can’t rate the game. Then, post, to the
current thread, game’s official name, and the players. Post
player’s Apolyton nicknames ONLY. For example, entries like Chairman
Yangestro or Prime Function Aki-Brian will not be accepted. Post the Apolyton
login. Players that are unrated should appear in parentheses.>>[/b]

> >[/b]

Q: When do you adjust
the ratings?>>[/i]

A: When a game enters
MY. ***0, e.g. 2110 or 2350, someone should post AC scores of all the players
to the thread. Then, depending on the scores, calculations are made, and I
adjust the ratings.>>[/b]

> >[/b]

Q: Where can I find
more information on this?>>[/i]

A: style='mso-spacerun:yes'> There’s a thread called
“Ladders” by Solver in the AC-Multiplaying forum. Find it, and read
it for more information on foundation of the system.>>[/b]

> >[/b]

Q: Have you done
something like this before?>>[/i]

A: Yes. This system
has been first implemented within Call to Power, and at the moment it’s
very popular, as we have 40 rated players there. Also, MarkG posts the PBEM
charts to the CtP front page, so the idea of doing this now isn’t new.>>[/b]

> >[/b]

Q: Who is doing all
the stuff here?>>[/i]

A: style='mso-spacerun:yes'> Here, in SMAC, I (Solver) am the only to do
that. I read the scores, calculate the new ratings and post them. However,
Keygen provides me with lots of help here, and the first two rated games came
because of his initiative.[/b]>>

For more on the rating system, refer to the Ladders topic by Solver in the AC-Multiplaying forum. And, here are some of the fundmental rules. All the rules are in one of my posts at Ladders, but I will include the rules here as well over time.

Initial ratings for the players: this depends on the highest SMAC/X AI level ever defeated by the player as follows: Citizen = 14, Specialist = 16, Talent - 17, Librarian = 18, Thinker = 19, Transcend = 21.

Delinquency rule: If a player doesn't play his turn in 24 hours, he is delinquent, and gets a next-turn from the previous player (probably CMN), to send the turn to the next one. Each time a delinquency occurs, the player loses 0.1 (one tenth) rating points. If there's explanation present either on the forum or in my mailbox, no penalty is applied. Explanation include things like lost Internet connection, Windows problems, etc. As in CtP, I have no time to monitor all of the rated games, so please, when a player is delinquent either post the fact at the respective ratings thread, or email that to me. I Know this sounds terrible, but as experience shows, it ain't that hard .

Elimination Rule:
Idea by Quinns.
If a player is:
a) Eliminated
b) Changes status in the game from rated to unrated
c) Gets end-turned three times in a row with no explanation,
then each remaining player defeats him 3 times.

Transcend Rule:
If a player completes the Ascent to Transcendence sequence, he defeats each remaining player three times.

Special thanks go to MarkG, who was kind enough to post the first ten at the Apolyton SMAC front page.

Ratings updated 11/03/01.

------------------
Solver - http://www.aok.20m.com
[This message has been edited by Solver (edited March 15, 2001).]
[This message has been edited by Solver (edited March 16, 2001).]
[This message has been edited by Solver (edited March 16, 2001).]
[This message has been edited by Solver (edited March 16, 2001).]
[This message has been edited by Solver (edited March 23, 2001).]
[This message has been edited by Solver (edited March 29, 2001).]
[This message has been edited by Solver (edited March 29, 2001).]
[This message has been edited by Solver (edited April 07, 2001).]
[This message has been edited by Solver (edited April 15, 2001).]

[Keygen]

Faction Strife MY 2110 Score Table
----------------------------------

Solver: 9
Mark13: 7
Keygen: 6
Tau Ceti: 6
Misotu: 6
Big Canuk: 6
Rynn: 4

[Solver]

Ratings updated with Keygen's Faction Strife M.Y. 2110. Congratulations this time go to myself for taking over the first place. If in CtP I have to languish among the last, at least here I can get a resonable place .

------------------
Solver - http://www.aok.20m.com

[Keygen]

Solver, why have you placed Googlie twise?
Rynn is Googlie!

[Keygen]

And I see Misotu in a higher place than Mark13.
Didn't Mark13 had an initial rate of 21.000?

[Solver]

Oops...That's due to the ties there. Yes, Googlie should prolly have the same rating , but I will now see what's with Mark13 and Misotu. Actually, Mark13 was better than Misotu by the results, so his rating should go up more.

------------------
Solver - http://www.aok.20m.com

[mark13]

That can't be right, surely...Rynn loses over eight-tenths of a point, simply because he hasn't founded a couple of his bases yet? I find that ridiculous, to be honest with you....also, I love the way that even though I managed second place from seven, I have gone down....

Not to criticise or anything... (j/k)

[big_canuk]

Solver:

I can't figure out your system. Perhaps you could post the mathmatics of how you calculate the scores, so individuals can confirm their accuracy. It would also be nice to see the "history" of one's ratings.

I realize that it might be impossible to do in a similar way, but I post a link to how the ratings are calculated in the Chess Federation of Canada (CFC). You can also go back to the main page and browse around. If you chose ratings, and list all, you can click on a name and see his/her history.

http://www.chess.ca/ratings_system.htm

I also see a fundemental flaw with trying to rate a game such as smac(x), in progress, where the ultimate winner is really what matters (aside from just the fun of playing, of course). In a seven player game such as this, winning at the beginning may actually lessen your ultimate chance of victory, because (almost assuradly), the trailing players will form a coalition to topple the top dog. A better position may be 2nd or 3rd, with breakout possibilities closer to the endgame.

Don't get me wrong Solver, I'm not complaining. You are terrific to donate your time to this effort. Anything is better than nothing. But if this system gives an "incorrect" output, we do have Tau's system as an alternative. It is just based on completed tournament games, which gives less timely, but more accurate outcomes, because it is based on final results.

[Solver]

Big_Canuk,
No offence taken. First, everyone note that I will soon fix the table - I'll post then. Today. Next, go to the respective thread in CtP-Multiplaying, and you'll see the formulas.
Oh, and I already know that I am terrific .
Next, it's not me who invented the formulas, it's Quinns who developed the way the ratings will be calculated. Perhaps I'm doing something wrong, but what I can promise is that each time I do it, I do it just like I did the previous time. A very important sentence, that's why it's bold.
So, even if I do it wrong, I at least don't change the formulas from calc to calc!
And, here's one important thing about the ratings: your rating will really depend on other's ratings. E.G., you beat 3 people, but get beaten by 1. However, that one that beats you has a rating that is much lower than yours. So, the rating difference is greater, and the adjustements are greater as well.
Next - this time, there were 4 (!) players with the same score after the turn came to an end, so it's a very interesting cause. Rynn has lost to everyone in the points, so it's no surprise for him to go down. And, if you have envy on me being in the first place right now here, just go to the CtP-Multiplaying, and see what my position there is! I'm among the last there .
I think that in PBEM games, the fun of playing is much more important than who the real winner is, don't you? And here, the way we measure it is the score, not number of bases, or whatever. No doubt, we could create a 'Pop-boom' Ratings System, where players would earn points depending on how much population they have in all of their bases, taken togther. Or, a research rating, when we'd look at how quikcly can one get tehcs, or... the list can continue for a long time now.
The reason why I have got the score of 9 here is probably that I already have discovered a technology in the game, I think some of you haven't yet, right?
Before you have some more questions, please wait for a ratings correction , especially merging Googlie and Rynn.
And the last thing I want to say in this post - remember that even though I want to improve the ratings system and make it more popular than it is now, I am always open to critics, as they really improve the system more than anything else.

------------------
Solver - http://www.aok.20m.com

[Keygen]

Yeah, Solver, you still got it wrong.
Mark13 should take the second or third place on the ratings table.

[big_canuk]

According to the ratings calculations page (based on tennis, rather than chess ratings), you get .25 for each win of a player rated the same as you.
Therefore, assuming we all start at 21 (I know Keygen doesn't, but close enough):

Solver, 6 wins = +1.5
Mark13, 5 wins, 1 loss = +1
Middle4, 1 win, 2 losses = -.25
Rynn/Goog, 6 losses - 1.5

Net = (+1.5 +1 -(4x.25) -1.5) = 0 great!

Right? Wrong? Did I understand this right? Or are the scores applied iteratively?
[This message has been edited by big_canuk (edited March 12, 2001).]

[Solver]

Ratings corrected (yees!). Now, I want to say a few words about Tau Ceti's system. Such a system is worth of respect, as he, just as many people put efforts in SMAC, especially its multiplayer part. But, I think that the main advantage of my system is that it offers more dynamical changes, that get even more dynamical as there are more rated games under way. So, if players are really paying attention to the system, this lets it get quite interesting, at least I hope that's what they think of the system.
Tau Ceti, if you happen to read this, please take a sign of my deep respect towards your system and the effort you have put in in such a game that SMAC is, to make it's PBEM better - my thanks! I don't view you as a concurrent by any means; I view you as a worthy colleague.

------------------
Solver - http://www.aok.20m.com

[big_canuk]

It's still not fixed. Googlie/Rynn has been fixed, but Mark13 is still lower than the 4 of us, not higher. The 4 of us were beaten by 2 and lost 1, so we should go down, I think (all initial at 21, except Keygen.)

[Rynn:]

All I know is that I'm avoiding 666 like the Devil



Sister Rynn

[Solver]

I'm not at home now to see, but it looks to me like Mark13 is all right. He had those wons, and during the calcualtions, at one moment his rating was better than mine rating. And then the last calculation step, "Solver beats Mark13" came. At the moment, Mark'u rðmzol gðu pjmmjr mdðo âzoj, uç zm gðu eztj mdj eçgjr rðmjs (Uçekjr) pjðmu dzldjr rðmjs (Âðrt), ðos mdnu dj îðâj sçgo âçrj mdðo Z gjom nè. Uçrrv lnvu, pnm Z gzee, çi îçnruj, rjîdjît.
JszmC Z ðâ mrvzol mç îdðolj mdj âjuuðlj, pnm uçâj èrçpejâu gzmd mdj eçîðe uçîzol ujjâ mç mðtj èeðîj ðîmzkjev. Z gzee mrv mç izb mdjâ, zi ðovçoj îðo rjðs mdzu.

------------------
Solver - http://www.aok.20m.com
[This message has been edited by Solver (edited March 13, 2001).]