Ladders

[mark13]

I would fully concur with Googlie's post - it hit the nail on the head. The internal scoring system is certainly not the answer - if someone could come up with a decent system that calcuated more truthfully a player's position in the game, I would certainly be interested. But not me.

[quinns]

Great explanation about the Statistics in SMAC!

I wonder, though, if the same thing happens in SMAC that happened in CTP. That is, in CTP, (sorry, it's the only thing I know, right now), there is a "Civ Score" and a "Power Graph". These do NOT track the same thing. Though there are similarities between the two, they are really quite different. One tracks "POWER" and one tracks "CIVILIZATION". Someone could have a very unhappy civilization but be very powerful. Someone could Nuke the rest of the world and "win" the game and be the most "powerful", but they would be considered a bit "uncivilized" (and not necessarily a very good player, either).

My question is, if this might be happening between your "F8 Key Stats Screen" and your "Power Graph". Are you sure that they are tracking the same thing?

You know, even if you were to come up with a really good algorithm to measure a player's ability at a given time, I THINK that it would be too complex to really use in a practical sense. I've been wrong many times before, but I can't imagine a truly workable while, at the same time, COMPLETELY accurate system for measuring a player's ability.

[This message has been edited by quinns (edited December 15, 2000).]

[Googlie]

quinns:

You are right, (and I erred in the previous post).

That formula above is the formula for determining the might of a faction as it reports in the power graph turn by turn.

The AC score formula (which appears when you hit the F8 screen, and at the end of the game) is:



#ADVCONCEPT4
Your Alpha Centauri Score is computed as follows:

(1) 1 point for each citizen of each base.
(2) If you have won a $LINK [Diplomatic=10005] or $LINK [Economic=10008] victory, score:
(a) 1 point for each citizen of a $LINK [Pact Brother's=3] base.
(b) 1/2 point for each citizen of any other faction's base.
(3) 1 point for each citizen of a surrendered base.
(4) 1 point for each unit of $LINK [commerce=29] your bases are receiving.
(5) 1 point for each $LINK [technology=140062] discovered.
(6) 10 points for each $LINK [Transcendent Thought=140088] advance.
(7) 25 points for each $LINK [Secret Project=110102].
(8) A $LINK [Victory=30] bonus if you have won the game.

There are then additional bonuses for early victory, using 'ironman' rules (essentially can't save and reload without a lot of hassle - guards (imperfectly) against someone replaying a move when they have hit a poor result in pod-popping or mindworm battling).

Many players steadfastly refuse to play 'Ironman' as the game has ben known to be unstable with .exe crashes, and who wants to have played 30 or 40 turns and not be able to restart on the last turn. (shouldn't apply to PBEM's though, as each turn is completed and saved anyway)

Anyway, I think these give a flavor of the magnitude of the task in devising an equitable running ranking system - and which gave rise to the "yawn" reflex.

To make it work in smac/x you ptobably need to link up with someone who has played extensively and who has broken down the formulae (like MariOne, Simpson II or maybe Velociryx) but their absence from this thread prolly means that they have little interest.

But if a formula can be created that meets us "whiners" criteria, I'm sure interest will perk up.



Googlie

[WhiteElephants]

It seems a mixture of both of these graphs/charts/scores would yeild an accurate result with a minimal amount of "new" formulas. If we can get both scores in numeric values maybe we could add them and divide by two to get an average of both. But, as Googlie pointed out, there are various ways to inflate the scores.

I can't quite conceive of anyone actually wanting to receive (x) amount of emails with one or even two (screen shots? or just the save games?) attached, opening them, computing the results, and then posting them weekly. Although I can't quite believe I'd actually want to, or remember to, send an extra email ever ten turns when I can't even work up the energy to post here after every turn.

Quinn, Solver, do you two know what your getting yourselves into?

[mark13]

Apparently it's worked with CTP, so I can't imagine that being a problem - still, I too would have reservations about this....it would seem a lot of hassle just to get a little chart up at the end of every week....still, if people are prepared to do it, I can't see it being a bad idea....

[quinns]

White Elephants -- Read my previous post. It is exactly this "hassle" that I'm trying to avoid! I think the simpler the better. As I implied earlier, anything more than ONE snapshot of a Power Graph for each TEN turns per GAME, would cause this thing to fail miserably. I, for one, sure won't go opening each player's game and record the results, divide by two, etc., then enter them into the system, even if it was only once every ten turns.

But I would be happy to do it as stated, (one Power Graph every ten turns), provided somebody from SMAC acts as a Ladder Administrator and sends me the data on a weekly basis (see earlier quinns post in this thread). It really doesn't take too much time (1 hour per week, about) now that the system is up and running.

Is the Power Graph really that bad in SMAC? Can someone really abuse it without compromising themselves later in the game?

Quinns -- CTP PBEM Ratings Processor

[WhiteElephants]

Just out of curiousity, how long do CtP PBEM/IP games take in terms of turns completed?

[quinns]

White Elephants - Regarding length of CTP Games - Well, I know of one game that is well over 200 turns right now. I think the only reason a game really "ends" is because players lose interest, or one player gets a substantial lead and it becomes just a long, tedious mop up job to "conquer" the rest of the civilizations.

One of the good points of the rating system is that it actually encourages players to stay in the game, and encourages the game leader "not to destroy" the rest of the nations, but just to stay in the lead (but not by too much) so that the leader's ratings will continue to increase at the ten turn mark and the other players will stay in, in the hopes of gaining the lead themselves, while still gaining ratings points for being ahead of the others in the game. For once the game ends, the ratings no longer change. There is no "bonus" for ending the game, (just as there is no bonus for ending the "real" world )

[quinns]

Well, I see that there are many more pressing issues in SMAC PBEM, right now, with the "flaw" in unfair difficulty level problems that I've read about.

But I hope this idea of "ratings and rankings" isn't dead here in SMAC. I really think most of you would enjoy this type of system.
[This message has been edited by quinns (edited December 19, 2000).]

[Solver]

Quinns, we've already been told that they might like the idea. But, they don't like the way we are going to realize the system. On the one hand, we shold now stop this, and this threead will probably die then.
On the other hand, I yet want SMAC PBEM giants to reply here, most of all I'd enjoy the opinion of Tau Ceti.
Quinns, I hope you see a SMAC box under your Chritsmas tree, so that you get to play it a bit, and understand how it goes. It would be quite hard to do something here, as long as you have almost no understanding about SMAC, and compltetly no understanding about the differences between SMAC and CtP.

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

[quinns]

Granted. Good point Solver. I would have bought Alpha Centauri a few weeks ago, but my wife told me not to buy any games because Christmas was so close (implying that she already bought "something" for me). Maybe that was just her way of keeping me from buying another $#@! COMPUTER GAME! (for awhile anyway ).

Would you like me to send a personal e-mail to Tau Ceti to get his possible direct feedback? He seems to be the main administrator of Alpha Centauri PBEM. It might be better if you did, due to my lack of knowledge regarding SMAC, but I wouldn't mind.

quinns

[Tau Ceti]

I echo many of the statements made by other players on this thread. I suppose the acid test is whether or not there is any interest in the idea. So far people seem to be saying "maybe, if you can work out a reasonable system of gauging performance". So, can you?

Personally, I am not convinced. The powergraph idea will not work - it is hugely biased in favour of tech early on and population later. It cannot be used as an objective measure. AC score is slightly more accurate, but only slightly. There are a great number of cases where the faction I would judge to be in the best position is behind in both power and score.

The root of this problem is probably the fact that these methods of measurement count only what you have right now, not your future potential, which is usually more important. Having a huge population is good, but someone with a large territory, many cities or weak neighbours ripe for plunder and capture is probably better off. It is basically good to have many techs, but it is more important to have the right techs. Getting lots of AC points for commerce is nice, but of little use if half of it is lost to inefficiency.

If a large and powerful faction falls to a coalition of lesser forces, was it truly that great?

In my opinion, the only reasonably objective measure of a player's skill is his victory statistic. Having a lot of assets is only useful insofar as it advances your primary goal of winning the game. If someone sneaks a victory out from under the nose of a faction that is more powerful on paper through the use of cunning diplomacy and sneaky tricks, that is just as worthy a victory as beating everyone into submission by being larger, and they should not be judged differently. Hence my tournament scoring system depends only on victory.

From the above it is probably clear that I personally would not be interested in participating, and I do not intend to substitute the ladder for the tournament score system, nor have the tournament officially support it. But if you go ahead with it and the players of a tournament game decide they want the game ladder rated, then why not? It pleases both worlds.

[Misotu]

You know, I thought this thread was tracking a new PBEM called "Ladders" so I've only just discovered what it's about this morning. Hence the late contribution, sorry

I second Tau's comments here - he has summed up the main difficulties with the Ladders system well. I've seen a number of games won by a player who was behind on the power graph right the way through most of the game - but despite what the graph said, the outcome was clear to all the players concerned well before the end. I have played both Civ and CTP - SMAC is very different, much more complex I don't see the Ladder as a replacement for the current tourny scoring system - that should stand as it is.

Having said that, you guys deserve a lot of credit for being willing to put time into a service that might add some interest, despite all the difficulties. Particularly your new proposal, where you're looking at AC score (probably, rather than power graph) at the end of the game, rather than trying to do a snapshot thing on current games.

There are still many accuracy problems - for example, in random map tourny games it's possible for one player to submit to another. This submission is generally an admission of defeat, but of course after a submission to the winning player the submissive's AC score could end up being rather good

And then of course, there are the team games ...

But still, absolute accuracy in any scoring system is a hard thing to achieve.

I don't know - if the Ladder were run separately from the tourny, and included other games too, it could add a bit of fun and interest to the SMAC/X MP arena? It might be so wildly inaccurate that it's really just a bit of fun. And then again, it might provide some useful help in matching players of similar abilities ...

I'd be willing to participate in a trial, to see how it panned out and whether it was popular. I don't really see what we have to lose?

- "Giant" Mis ( Thanks Googlie )

[Misotu]

PS I don't know if it's possible to rename a thread, but one reason for the lack of responses may be that others have assumed, like I did, that this is a turn-tracking thread for a PBEM called Ladders.

If it's not possible to rename the thread, you might consider starting a new one with a link back to the discussion so far?

Just a thought ...

[quinns]

That's great Misotu! All we need to do for a trial is rate any one or two games that you are currently playing by getting a concensus among the players (that the game will be rated). (We could even rate the "already completed" games but I doubt that the losing players would consent to the rating status (post mortum) )

Right, good suggestion about renaming this thread to something more descriptive. I think Solver has some ties to the Apolyton Administrators so maybe he could rename this thread. Any suggestions? "Possible New SMAC Rating System" ???

[quinns]

Thanks Tau Ceti for your comments and concerns. Solver and I greatly appreciate your opinions on this.

Aha! I might have a compromise to make this CTP Rating system work with SMAC, with some modifications as to WHEN the ratings are recalculated.

Tau Ceti, "Game Completion" was the original method that was going to be used for Call to Power regarding ratings determination. Many players voiced their opinion that CTP games rarely "complete" and just sort of fade out, while other games go on forever (years and years). We needed to have a method for scoring that would update the ratings more dynamically. So we came up with the "every ten turn" results method (which seems to work quite well in CTP PBEM).

This does not seem to be true in SMAC PBEM. It appears that games complete quite frequently. This, then, would lend itself to use the Ratings System for SMAC based upon Finishing Order at Game Completion . I see that you have five month "seasons" in SMAC tournament rankings. This could, of course, still be done, but simultaneously, we could also start to track the tournament matches' (and other matches') finishing orders within the Ratings System (just to see if there was any interest).

It really would be no problem for me to do this. All I would first need would be the initial ratings of all players that wish to be rated. We determine INITIAL player ratings in CTP PBEM by giving the following ratings for EVER defeating the Artificial Intelligence (solo): Chieftain = 14.000; Warlord = 16.000; Prince = 17.000; King = 18.000; Emperor = 19.000; and Deity = 21.000. I would guess that SMAC has somewhat similar difficulty levels that we could extrapolate to use in a similar context.

After that, the player ratings would change depending upon the finishing order of each of the games and the player ratings of the players involved. For example, in a 6 player game there would be 15 "results". Finishing player 1 defeats players 2, 3, 4, 5, and 6 -- finishing player 2 defeats players 3, 4, 5, and 6 ... etc. for a total of 15 results in a six player game.

The exact formulae used to determine the change amount and examples can be found under the CTP Multiplayer forum thread entitled PBEM Ladders.

If there is any interest in this, I would be happy to help out.

Quinns


[This message has been edited by quinns (edited December 19, 2000).]

[Misotu]

Yeah - or New PBEM League Table? or ... well, anything would do as long as it's clear it's a discussion thread not a turn-tracking thread. In fact - how about "Ideas for PBEM League Table - Discussion Thread"

[Solver]

Now here go my comments about the things going on:
First of all, thanks PBEM Giant Tau Ceti for replyuing to this thread. And yes, I know that the future potencial is important - but if you coem on and improve, the amount of your AC points will be yet higher. And this will lead you to win in the ratings. It might be so that the one who is down the PowerGraph wins the game (while this doesn't happen very often in CtP), but the one who takes the 1st line there, yet got it more or less deserved, he did something to get there.
That's why I want to use the AC Score instead of the powergraph here. When the turn x9 is going, everyone emails us with their points, then quinns can recalculate the ratings just the way he did in CTP.
Game Completion was the original thought in CTP PBEM because I first got the idea of Ratings before I started playing PBEM - thus I didn't know it takes them so long to end. That's why we came up with the idea of changing the ratings each 20 turns - resulting in the each 10 turns change.
Quinns, I do not yet see how can the Game Order by Completion in SMAC lead to dynamical rating changes. What do I mean by dynamcial changes? It's when, updating once a week with 5 games, we have something to change.
This hasn't been yet brought up, but I believe that the idea of taking off rating points for delinquency will also come true here, if the rating system is used. In CtP, this idea came up after seeing how slow some players tend to be.
The difficulty levels in SMAC are quite close to CTP, but I can't name them just now - I only remeber it starts with Citizen, ends with Transcend and that Talent is the third, followed by Librarian. I think that the levels under Librarian are very easy to beat, but here it's not the question being discussed. We could yet give 14 points for Citizen, 16 for the next, and so on, like in CTP.
Misotu and quinns, yes, I have contacted the Administrators of Apolyton in order of renaiming this thread. In case this is possible, it will be done, if not, I'll start a new one with a link to this thread in the first post.
Yes, I am excited about the idea of PBEM Ratings trial in SMAC. This means that the idea is not dead here, it means we keep going on with something, and that it might someday come through. In CtP, it took us about a month from my first post about it, to make it something sensible and working. That's why I didn't at all loose heart hearing what the first replies, by Googlie and WhiteElephants, were on this thread. Now, in the SMAC ladders, we've gone farther then we ever were, and hopes this comes through.
Oh, and then I will no longer be entitled CTP PBEM Ladder Administrator, I'll be PBEM Ladders Administrator .

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

[Googlie]

Solver - you've identified the major flaw in SMAC/X as a game.

Unlike Civ, CivII, CtP, CtPII and ToT, all of which require the winning player to build a victory, using multiple bases for a quick build or an extended build in a single base, in the smac games a player can lurk in the weeds and then buy a victory from under the noses of the leading players.

When one player has built the Voice of Planet, all players can then commence building teh Ascent to transcendence.

Based on turn order, if the Voice completer doesn't immediately build the Ascent, it's open to the next player in the turn to do so, and so on.

A player that might be 7th in the power chart, and hopelessly behind, can buy his way to victory by cashing in crawlers, disbanding base facilities and disbanding units, and no matter if all the next five players do the same thing, next turn he wins (It's too late for player #1 to do anything about it - the chance has gone previous turn)

When smac fist came out, the old OWO boards were full of posts about how this was such an unsatisfying experience compared to the CivII one.

(Many of us in single player games have partially solved this by tweaking the alpha.txt to set the cost of the ascent at ten times it's value - 2000 instead of 200 - which makes it nearly impossible to buy in one turn. Minerals need to be crawled in from other bases, and the end game lasts some 30 to 40 turns longer, necessitating those advanced weaponry units to be built to stop the AI from PBing your building base.

Totally different game then.

But I digress.

I don't necessarily agree with Tau on this one that who's won is really the only determining factor in play. Maybe a combination of an aggregate accumulating score plus a hefty win bonus is the answer.

But the more we keep this thread ongoing and open for discussion the more these minds here will arrive at an equitable methodology.

I'm warming to the idea ........



Googlie

[quinns]

We "could" possibly use Solver's method, (i.e. screen-shots of AC Score display e-mailed every 10 turns on the ...9th turns, e.g. 9,19,29,39,... etc.[provided Solver does the pre-sorting of the raw data for me ] ), while at the same time giving a substantial bonus for defeating players, via: 1) the game officially ending; 2) annihilation; 3) resignation; or 4) inactivity (i.e. "next turned" three times in a row). We are currently proposing (in CTP) that the elimination bonus be a factor of 3 times the normal ten-turn rating adjustment for "eliminating" another player.

Just a thought.

[Solver]

Quinns, as for the AC score, I wouldn't even like to have snapshots of it, just to get the number in email. If a game has five players, then I would have to see five screenshots, as you only can see your own score (hey SMACers, am I right here?), so this would be painful.
As for our elimination bonus in CtP, none of the rated games have yet been with eliminated players. BTW, I haven't seen this posted, but I like it...
Back to the SMAC idea, here's an explanation for you of what Googlie meant by buying the victory:
SMAC has several types of victory. One is Conquest when you eliminate all the others. If you have pact with someone (like Alliance in CtP), you don't have to kill him in order to be a winner.
Second is Diplomatic (that's what I did always easily achieve). There's Planetary Council, and each faction has votes in it. Basically, it's easy to get over 50% of the Council Votes. Then, you can just vote at a speciffic issue, and if you win there, you win the game. I often got over 50% votes myself, in other cases I was able to make the AI factions vote for me by giving them lots of money.
The third one is Economic victory. It's hard to achieve, and it's not quite often to see.
Fourth one is Transcend victory. You achieve it by making your humans transcend. To do it, you must build Secret Projects (wonder in CtP). The first one is Voice of Planet - you need almost all the advances to build it. After it's build, you need Ascent of Transedence. All civs can build the Ascent, when the Voice of Planet has been built, no matter own they the neccessary techs or not.
So, if you have got enough energy (money), you can just rush buy the Ascent Secret Project, and that's at it.
Yes, this might be a real problem here, but I yet believe it's possible to do something with it. What? Noone knows .
And Googlie, nice to see that you no longer criticise us the way you did when I first posted the thing.
Note: since I have sent MarkG a message, he hasn't checked it, let's wait a bit more.
And quinns, please, all in all, GET ICQ! I finally want to contact you in a good online chat someday, and anyway that's a nice way to stay in touch.
OK, that's all for today from me, will check back tomorrow.

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

[Misotu]

I think one of the more attractive features about your system is that it does distinguish between 1st through up to 7 places. Using the AC score is not a perfect way to measure this, but I think the idea of aggregates plus a hefty win bonus would probably be a reasonable starting point ...

[Misotu]

I just thought ... forgive me if this is a stupid question demonstrating that I haven't understood a word of the discussion so far ... but would this system simply favour the people who play a lot of games? As in, the more you play the more points you get no matter how well you do in terms of result?

[quinns]

Misotu, good question.

No, the system doesn't favor those who just play a bunch of games. It was designed with this in mind. There is a small penalty for inactivity within the ratings (-.05 ratings points per month of rating inactivity), but that's it.

You gain points for winning and lose points for losing. However, the gain/loss becomes more and more minimal depending upon the DIFFERENCE IN LEVELS of the players involved. A person rated 21 who defeats a person rated 16 only gains about 2/100ths (0.02) of a point, (and the 16 rated player would only lose 2/100ths of a point). Whereas, if the 16 rated player defeats the 21 rated player, there is, essentially a "cap" of 5/10ths (0.5) of a ratings point gain or loss, (the formulas handle all the details).
[This message has been edited by quinns (edited December 20, 2000).]

[quinns]

Happy Holidays to all you SMAC PBEM'ers!

Merry Christmas and Happy Chanukah to everyone!

[Solver]

Hehehe, you wish everyone Happy Chanukah. The Jewish are everywhere .
And, keep figthing for the ladder system while I'm out of town.

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

[DilithiumDad]

I would like to endorse the idea of 10-move games --as a children's party game! I set up a hotseat game at my son's b'day party as one of the "carnival of games". The game was won by an 11-year-old who had never played before (he was Morgan and so got bonus points for commerce). My son and I were 3rd and 4th. So it's agreat kid's party game by has nothing to do with playing skill!!

------------------
Creator of the Ultimate Builder Map, based on the Huge Map of Planet

[quinns]

Thanks for the endorsement Dilithium Dad. I'm not quite sure what you're really saying, but I believe you are being sarcastic in your remarks. Well I understand what you are saying then, (I think). You believe that even a complete novice could be winning at the ten turn mark due to the complete "luck of the draw". So that the only fair measure of who is the "best" player is the one who wins the game. Is that right?

My only remark to this is that I have seen the same thing happen in Chess, (just about the "fairest" game in existence). I saw a complete Novice beat a Grand Master out of pure luck. The Novice got "lucky" when the Grand Master became careless and lost his queen immediately. Does that make Chess a party game for children only? (What's wrong with children being part of the ratings anyway??) No, of course not. Over time, the Grand Master would win many more games than the Novice.

Does this sarcasm match what everyone else feels here? If so, then I don't have a problem with using "game completion" as the measure of changing the ratings.

I think Solver was only attempting to find some way of causing the ratings to change on a more dynamic basis, (i.e. every 10 or 20 turns). About how long does the average SMAC PBEM take? If it is more than two months, then SMAC ratings may become stagnant and people may lose interest in the rating system. The ten-turn method used in "Civilization Call to Power" PBEM has seemed to promote quite a bit of interest and increased turn around times. But it may not work the same way in SMAC.

[Keygen]

This topic is getting big enough as I see .
Hi Quinns, I think it's time to express my opinion about the hole idea .
As far as I know the best evidence for somebody's skills is how many times he can win a game.
A player can build a legend around his name by winning as many games as possible and more important by lossing as few as he can. Of cource those wins must be done against as many different opponents as possible (you should consider about it in the ratings).
Most of the PBEM players continue to play for a long time games via email because of the great fun and pleasure that only human players can give .
But the main reason that people begin PBEM games and one of the most important reasons to continue playing PBEM games is to compete with human opponents and beat them.
Win is a very important issue and is what gives the most satisfaction to a player.

You should probably thinking right now that I am against the whole idea of the rating system that both Quinns and Solver are proposing to the SMAC/X multiplayer community.
WRONG! I think that it's a very good idea. It takes a long time to complete a PBEM game so it takes a long time to satisfy yourself with a winning and far more time to satisfy (maybe ) yourself by winning the revanche of the game you just lost!!!
With the rating system we have the chance to compete with the others not on winning only but on best management of factions.
Yes, we can very frequently monitor our (and other's) progress on managing our faction and be compared with others on areas such as Technology, Economy, Military, etc.

My opinion is that both Tau Ceti's and Quinns-Solver's rating systems sould coexist.
The first for showing the (really) best players and the second to keep track on a regular basis to other importand fields such as managing factions.

I would like both of my games "Keygen's Classic" & "Keygen's Alien Crossfire" to be the first rated PBEM games by Quinns & Solver's rating system if the players that are currently playing are interested of cource (I am in ).

Go for it Quinns & Solver!!!

[quinns]

Alright Keygen!!

Good piece! As soon as the other players in your "SMAC Rated" games agree to the rating system, I'll need to get their initial rating. I think you are familiar with this "initial rating" in CTP PBEM, so maybe you could come up with an equivalent way of getting their initial ratings in SMAC. I should be getting SMAC in two days and I might have some novice-like advice then regarding initial ratings.

Thank you very much for your input!