Request for comments about the number of the winners for LT34

[wieder]

With Lt33 there has been a problem with the new players not being able to have a realistic possibility for winning the game because a similar sized alliance with experienced players was able to beat a group of not that experienced players.

http://www.longturn.org/ranking/

One idea to solve this problem would be allowing the new players to form bigger alliances and that way gaining more chances to actually win the game. The initial idea was to make winners of the last n games (for example last three) to be counted as 2 players. This however would have some issues because good players may have skipped several games and those who have lost several times would still be considered equal to those who may know how to play but have only played once but yet lost.

Using the ranking scores can be used to give abetter idea about how well the players can play the game. However because the ranking score can vary too much, they should be filtered. One way to do this would be using square roots for the scores. The best ranking score would give a sqrt(3630) = 60,25 and the lowest sqrt(194) = 13,93. Those players who are new to the game will have an initial ranking score of sqrt(1000) = 31,62 and this could be used as a base value for everyone without a ranking.

We could limit the number of people in one alliance to having max of 200 filtered points. This would allow 6 completely new players to ally or (currently) up to 8 low ranking players to form an alliance. In comparison the top players could form an alliance of just three players and they might be able to have one more player with a ranking lower than the base score of sqrt(1000) = 31,62.

This approach would encourage allying with new players or those who have lost some earlier games.

New players will have a ranking only after 2 games. Because of this we should decide if those who are playing the second game would have a synthetic ranking score (lower/higher than the base score of sqrt(1000) = 31,62 or if the base score of sqrt(1000) = 31,62 should be used anyway.

Please comment on this and feel free to say if this sounds too restricting or if it's still not enough. You can also suggest another way for filetering the scores or even a better way of giving ally points to the players.

[cgalik]

I agree with the idea in general.

But I didn't take the time to follow your math.

[StratThinker]

Players who have a higher ranking than what they should have will be avoided, while everyone would want to ally with someone with a ranking lower than what they should have.

In the best case this would be self correcting; players that are ranked too lowly can be in larger alliances and thus would win more often, which would increase their ranking to what it should be. While the opposite would happen to players ranked too highly. In the worst case it could cause the player who are ranked too highly not to play.

[mmm2]

I like this analogy: If you had 20 mice fighting against a cat they might be able to kill the cat with their small bites, but that still doens't take away from their victory just because they have big alliance. But I disagree with using the ranking scale to determine this. I think you could try to make such report about team strength each turn, and that might help make the game more enjoyable to track progress, but for making algorithm for winning, don't bother, because I think most players aren't so concerned about win/loss after game - once game is over it's over. You want to create tools and reports, etc that assist for in-game strategy. One that I mentioned in the past that would be really good would be to make report showing list of shared vision webs with total points and whom is in each shared vision web. That can easily show that even if there is "BIG" team of 45 noobs that they are actually weaker than the team of 5

[wieder]

Great comments.

It's of course true that players with too high rankings would not be picked that easily and the other way around. Then again this is more for limiting the team sizes when there are too many good or great players in a team and since it looks like this is not going to be a team game, location is still the most important factor when picking at least the initial allies. No one is able to pick and succeed on picking the n lowest ranking players if they are all over the map. You can pick some far away players but if all of them were far away from each other, that would also give some advantage for the enemy.

The top 20

1 kryon (3630)=60,25
2 bli (3106)=55,73
3 edrim (2975)=54,54
4 terror (2938)=54,20
5 kevin551 (2928)=54,11
6 mrsynical (2740)=52,35
7 wieder (2398)=48,97
8 akfaew (2248)=47,41
9 duncan shriek (1975)=44,44
10 dimitril (1964)=44,32
11 joris (1964)=44,32
12 kull (1964)=44,32
13 det0r (1777)=42,15
14 elrik (1748)=41,81
15 ollikka (1733)=41,63
16 modeemirotta (1733)=41,63
17 jhh (1560)=39,50
18 jumangee (1540)=39,24
19 oletraveler (1508)=38,83
20 mmm (1471)=38,35


More scores

(1200)=34,64
(1000)=31,62
(800)=28,28
(700)=26,46
(600)=24,49

There are really 3 things for this matter.

1) Can a system like this or a modified system like this used for a ranking game?
2) Should we make it possible to have more players in a team if they are less experienced or something like that?
3) If 1 and 2 then what would be the algorithm to give the points for the players? There was 2 ways for doing that in the opening post and while no system can be perfect, it would be good enough for LT34 if we can find one that's better than the current one with a fixed number of winners.

[Nimrod]

This problem is made all the more simpler if LT34 is a Team based game, with a good mix of new & veteran players within each team. Therefore veterans must shadow new players and take them on as "apprentices" if they want to win; new players get a chance to learn a lot; new players have a reasonable chance of winning.

I'd leave the final decision as to team makeup to Wieder, but obviously existing ranking should be used to determine "veterans" and figure out which ones will be on which team, for instance.

[mrsynical]

I don't know details about LT33 but it seems too complex to try and force people to pick alliances with "weaker" players. I would suggest trying to minimise "out of game rules" as people just find ways around the rules - e.g., do large alliance with good player, once everybody is dead, some players in the alliance resign, then smaller alliance declares "full" victory.

LW sorts out issues of grouping players of similar skill levels...

[wieder]

At the moment this seems to be to complex. Maybe we can return to this topic once LT34 is over and speculate with "what if" scenario.

[mmm2]

i expect there will be a massive overhaul to the unit, tile, wonder, etc properties in LT34... And also if they are able to test out 2.5x in LTex and then use it in LT34, it's going to be massively different game.. even veteran players will have a somewhat steep learning curve - that's going to level the playing field anyway..

[Nimrod]

i expect there will be a massive overhaul to the unit, tile, wonder, etc properties in LT34... And also if they are able to test out 2.5x in LTex and then use it in LT34, it's going to be massively different game.. even veteran players will have a somewhat steep learning curve - that's going to level the playing field anyway..

I'd venture to say that the truly veteran players keep up with rules changes and ruleset changes regardless of the sheer number of changes, and are certainly monitoring both this forum and Trello with much interest. So therefore, the newer players are actually disadvantaged even more so.

[kevin551]

And also if they are able to test out 2.5x in LTex and then use it in LT34, it's going to be massively different game.

In fact the ltex25 ruleset is intended to be the same as the recent longturn games. The one new feature which I want to include is finite moves on rails. I think this will help rather than hinder the novice players.

[evan]

Sorry Wieder,
i was trying to find where it was written about the number of winners - it's not finalised yet? Kryon was suggesting a higher possible number, /7 eg.

If it's not decided yet then i also think a higher number of players in the winning alliance is better.

Many of us are trying to keep out of the way, and not draw attention to ourselves. We're not as experienced as individual players, let alone working as a fighting team with a group of others.

I was worried at first that it would just be like the last game, but i think the mix of rule changes is quite good.
This would help as well.
The little countries might see the advantage in engaging in exactly the kind of early combat you're looking for.