PBEM tournament, looking for players!

[Hunkpapa]

This is my first posting ever in this forum. I am very interested in playing in a PBEM game of Alpha Centauri X-fire. I have never played in this format before but would like to get acquianted with it. So I will need some help setting it up and recieving/loading/sending turns b/c I really have no idea what to do. I will be able to do at least 1 turn/day. And would just like to start with one game for now.

email: pmarc_wcpa@msn.com
[This message has been edited by Hunkpapa (edited February 04, 2000).]

[zsozso]

I would like to reward faster victories more - although your point is good about being able to beat a weaker opponent faster, so opponent rank also has to be taken into account.

In my example, Y would get higher score for slower victory over the good player X, if we follow the scoring system you or Korn469 suggested and compute the scores at game completion based on rank. That was the argument against those - otherwise good - scoring systems. My conclusion is, that any rank-based score system has a time related problem with parallel games regardless of the fact that you determine the score at the start or at the end of the games.

However, there is one solution, which would overcome this problem: we have to use a rank based scoring system, which recomputes all the ranks and all game scores after every game completion. I.e. player A beating player B may reward him a certain score at the time when that game competed, but the score for that game may change in light of later games of both A and B, e.g. if player B wins many games later, then the score A gets for that earlier played game would increase!

I suggest a score-stealing system:

• All scores are re-computed after every game completion, even the scores gained in earlier completed games.
• The score computation is done in two phases: score awarding and score stealing.
• In first phase, each winner is awarded a certain score for each game won. E.g. it could be based on M.Y. of the victory, let's say (2600-M.Y.), i.e. each win is worth at least 100 points, theoretical maximum being 499 for winning in 2101, but practically you could expect around 200-400 points per win.
• Second phase is the stealing, where the awarded points of each winner are divided for all the games the player completed. E.g. if a player completed 5 games and won 2 of them, then the sum of the 2 awared points would be divided by 5. The player keeps 2/5 of his awarded points, but loses 3/5, 1/5 stolen by each player who beat him. On the other hand, he steals some points for the 2 games he won (from both losers there).

The stolen points are computed purely on awarded points, you do not lose again from stolen points, so it is not an iterative process, only 2 phase.

This way beating a good player who won (or going to win - order doesn't matter!) a lot of games is more beneficial, because you can steal a lot of points. On the other hand beating on newbies who do not win any games, will only grant you the awarded points, nothing is stolen.

I think this is a simple enough system, but still contains the ranking factor and also the faster win better factor. On the other hand, winners will want to play other winners to be able to steal more points, so we get the benefit of the level-based system. In fact, we could even introduce an extra rule to include the level system, i.e. games would be scheduled between players who won the same number of games so far.

Another advantage is, that players, who play a lot of games but only win a few, would not be scored above good players who play few but win all. E.g. one player who played 10 games and won 3, probably not better than one who won 2 out of 2. A simple level-based system would rank the 3-winner/7-loser player higher than the 2-winner player.

Any thoughts, counter-arguments ?
[This message has been edited by zsozso (edited February 04, 2000).]

[Pagan[CyC]]

Zsozso:

I was not explicit enough in my explanation of the scoring system I proposed; it would completely recalculate scores after every additional game completion.

[Paul]

zsozso, just let me see if I understand that scoring system of yours. For example, there are four players: A, B, C and D.

1st game: A, B, C -> A wins in 2300 -> A gets 300 points
2nd game: A, B, D -> A wins in 2360 -> A gets 240 points
3rd game: A, C, D -> D wins in 2420 -> D gets 180 points
4th game: B, C, D -> C wins in 2240 -> C gets 360 points

A has 540 points, won two out of three, so he keeps 360 points and the remaining 180 points are stolen by D who beat him in game 3.

B has 0 points, so there are no points to be stolen and he stays at 0.

C has 360 points, won one out of three, so he keeps 120 points and loses 240 points which are split between A and D who steal 120 points each.

D has 180 points, won one out of three, so he keeps 60 points and loses 120 points which are split between A and C who steal 60 points each.

This leads to the following final scores:

A: 360+120+60=540
B: 0
C: 120+60=180
D: 60+180+120=360

Did I get that right? Or did I misunderstand anything?

[zsozso]

Yes Paul, you got it all right. Your example also nicely illustrates how the stealing works to switch places between C and D.

C and D has beaten each other, so it is not clear which one better from that point. On the other hand, A was clearly the best and B is the worst player. The straight scores would've put C ahead of D due to its faster victory. However, C has beaten the weakest B, while D has beaten the champion A (even though in a longer battle), so D deserves the second place!

Pagan[CyC], sorry that I did not understand that in the first place. In that case, your system is in basic principle equivalent to this 'stealing' system, but yours goes further by iterating the process - we could try that and see how it works...

Zsozso
[This message has been edited by zsozso (edited February 04, 2000).]

[Paul]

Ok, then I think this would be a good system to calculate scores.

Also, I originally said I could play 3 games at the same time, but after playing some turns in my games I think that I could play a 4th game.

[MichaeltheGreat]

Interesting thoughts on scoring systems, but I don't think these games will take that long, depending on the players - with the map layout, accellerated start, and three player no-AI format, I have a hard time seeing these games go past 150 turns.

In my PBEM experience, I've found elite X Shard Fusion Choppers hard to resist - my longest running PBEM (with five plaers and 2 AI) is 155 turns, on a much larger map, and is 5-10 turns from completion, depending on my last victim's appetite to play to the bitter end.

On this map, with each island in PB range of it's neighbors, the rush to Spaceflight and MMI is going to be decisive. If players resign when their situation is hopeless, we should see most games resolved in the 3-6 month range.

[Genaciv]

MtG said:

quote: If players resign when their situation is hopeless, we should see most games resolved in the 3-6 month range.

Sooo.. are you saying that players SHOULD resign when things start looking.. bad?
If so, then that should practice what you preach and resign from our game here, and the CoC vs Germany game.. :P~~~

- General Lee Clueless (extremely clueless...)

[MichaeltheGreat]

Genaciv, since I'm a collateral descendent of Maj. Gen'l Fitz Lee (Rob't E. Lee's nephew), I'll invite you not to sully the Lee name, despite what Jimmytrick says.

You're right, I'm doing terribly in the CvG game. Industrial automation by 2116, a pair of 4-3-2 rovers on the way to visit Jimmytrick's 4 alien artifacts and their escorting scout patrols, and 2 of 3 enemy factions infiltrated and mapped. When I'm elected governor in the next turn or two, I won't need to waste a probe on infiltrating you. Since I've also got Cent. Empathy, I'm switching to a green economy so I can do 100% research, that way, I'll make sure that my needlejets are flying your way by MY 2150. Have a nice day.

[Genaciv]

DOH!

- General LEE Clueless

[LTEC!]

I'm available to play a couple of games. Factions preferred UoP-Hive-PK-Gaia map depending, of course.
My email is claudiam@bc.cablemas.com .



------------------
You can have your cake too...

[Genaciv]

LTEC! is just MtG's brain in disguise.. and not only that, but she is pure evil! Beware her dirty tricks, and legendary predispostion to BETRAYAL...

- Genaciv

[MichaeltheGreat]

LTEC! has her own brain. I'm just her "military advisor" when it comes to SMACing people around.

For Apolytonites who don't know, LETC! is my wife, and a pretty good MP SMACer in her own right. We have a LAN at home we use to play a lot of IP games on.

[Tau Ceti]

Is AXT004 ever going to start?

[MichaeltheGreat]

Sorry about the delay - I've had comp and R/L schedule problems the last two weeks, I'm working on getting it out late tonight. (GMT -8)