3 player game starting positions

[sparrowhawk]

It's become apparent that whereas the quadrant system for determining starting positions is fine for 2 and 4 player games, it is too inequitous for 3 player games.

So, should I implement "tiers" of bands that run horizontally or vertically (chosen randomly by system on game start up).

What happens if you are the piggy-in-the-middle though?

[Ringthane]

Is it very hard to implement 3p starters by 120º slices?

[sparrowhawk]

If you can give me an algorithm, then no ;D

Seriously though, it's not a bad idea - can anybody out there tell me how? Maths is not my strong point, let us put it that way

[Ringthane]

As far as I know, the map has a two-axis system with 0,0 at the NW-most square, right? Since this is only for initial placement purposes, is it possible to define the center at one of the middle squares (never mind which) and then use polar coordinates with a radius limit, less than the larger map diagonal, to delimit the slices?

[Matija]

4 p games are a problem even more often. As I mentioned it a couple of times before, I think starting positions are the part of the game that is the most in need of revising.

[sparrowhawk]

Can somone summarise the alternatives offered? I did suggest a completely random start position (as an option only), but this seemed to be met with little enthusiasm, so if we could list them all, then maybe we can find a solution.

Those that I can think of:

- Current Quadrants
- Smaller quadrants so that no one starts on the edges
- Tiered bands for 3 player games
- 120 degree bands (Ringy's proposal above)

Please add to this list. thanks

[Ringthane]

- Current Quadrants
- Smaller quadrants so that no one starts on the edges


- Tiered bands for 3 player games

- this might have the edge problem far more often than not... two of the three tiers would have three edges each!


- 120 degree bands (Ringy's proposal above):

- like a roulette, that is, for each player a "slice" of squares is selected based on three previously defined sets divided by radial coordinates; then from that slice, a square is selected. for the next player, the first slice does not count anymore.

[Matija]

The idea that I like the most is that a set of 100 'good' starting positions is made. One is randomly picked among them, and perhaps offset a few squares.

I think that would generate a sufficient appearance of randomness not to become repetitive.

[Matija]

Perhaps you could implement all the suggested methods, then, at game creation, pick one at random.

[Ringthane]

But Matija, what if all 4 players start within the max limit (say, 5 squares) of each other?!

[Matija]

What do you mean? I don't think that would be a valid starting position. If two players moved in the same direction, the second one would probably have no chance of winning the game, and we want to avoid situations like that.

[Ringthane]

Ah I see, you mean to pre-validate the entire thing as an array, not position by position.

[Matija]

Yes.

[celebaglar]

I have to admit I am extremely leery of pre-determined starting positions. Not only will they spoil the variety by ensuring "fairness", but I also believe that no matter how you disguise it, players will eventually suss out the patterns and this will have the effect of mass predictability. It's not so much a question of knowing exactly where your opponent starts as much as knowing for sure where he does NOT start.

I don't think completely random starts would help either, as they would lead to even more games rendered unplayable from the start.

IMHO, a true random start tempered by boundaries is the best compromise. That's what we have now, but the perhaps the rules for boundaries need tweaking somewhat.

How about this:

(a) Use Ringy's suggestion of equal sized pie slices to determine starting sectors. Three equal sectors for a three-player game, four for a four player game.

(b) Rotate the sectors around the center of the map, so that players don't know where the boundaries of their sectors are. This would need some fairly simple math to implement.

(c) Create three or four concentric "bands" around the center of the map, and ensure that all players start in the same band. This will eliminate the situation where some players start near the center and others near the edge. Since not all locations on each band are equally good, it would not affect the variety element too much.

[Ringthane]

c is too dangerous. box rotation seems good, whether by pie slices or quadrants.