Beginner's Guide to EA Arcoscephale

[lch]

Originally Posted by chrispedersen

Yes, you are pumping dominion - 10 Dom score, awake pretender, prophet etc.

But as the size of the map increases (r2) your dominion is only linear.

Strategies which rely on dominion therefor fare best in small to medium maps.

That sounds like you think that dominion spread is limited by max dominion. Is that the case? I thought that the Dominion only limits the maximum dominion in a province, but not the dom spread, which only depends on the amount of temple checks being generated.

[Sombre]

I believe lch is correct. However dom 10 is much more powerful on a smaller map, because no-one really wants to build lots of temples, which is what would be required on a larger one. You have your prophet, starting temple and god for free and they make a much bigger impact on a small map.

[chrispedersen]

Originally Posted by lch

That sounds like you think that dominion spread is limited by max dominion. Is that the case? I thought that the Dominion only limits the maximum dominion in a province, but not the dom spread, which only depends on the amount of temple checks being generated.

No, thats not what I'm thinking.

What I am saying is this:

Dominion will spread from your dominion soursces - pretender, home province, temples, etc.

As the radius of yoru empire expands, the likelihood of a dominion spread into an adjacent territory decrease by r2.
The number of dominion increases linearly, f(n) where n is a number of temples (with things like priests etc tossed in for good measure). You also have competing to dominion to consider as well.

However, my point is that your ability to project dominion over a significant percentage of the board decreases as the size of the board increases absent military (or other) actions.

[analytic_kernel]

I think your analysis of dominion spread has some merit, but it seems to be based on a fixed number of sources that do not scale with time (temple checks occurring at each turn) or space (more temples built as number of provinces increases). If you are building new temples, then the n in f(n) is not constant, but rather n(t). Also, irregular tile (province) geometry and variable numbers of neighbors per province make it more of a discrete or statistical problem that might have some difficulty being modeled with a simple polynomial function.

[MaxWilson]

Thanks, analytic_kernel. I had never before considered taking the topology of the map into consideration.

-Max

[chrispedersen]

Originally Posted by analytic_kernel

I think your analysis of dominion spread has some merit, but it seems to be based on a fixed number of sources that do not scale with time (temple checks occurring at each turn) or space (more temples built as number of provinces increases). If you are building new temples, then the n in f(n) is not constant, but rather n(t). Also, irregular tile (province) geometry and variable numbers of neighbors per province make it more of a discrete or statistical problem that might have some difficulty being modeled with a simple polynomial function.

Thanks. Obviously, its n(t). But that doesn't change the fact that is a lower order than r2 - which essentially is what my point was.

[analytic_kernel]

Chris, I hope I'm not being too anal[ytical] about this, but I think that r is r(t), and that we don't necessarily know the form of n(t) or r(t). Looking at your number of posts, I'm guessing that you are a seasoned player, so I trust your judgement. But, if we wanted to be semi-empirical about it, we could collect data (possibly with the --statfile option, though I haven't tried it) on number of provinces and dominion for each turn, and then fit the data sets as a function of t. This would give the forms for n(t) and r(t).

Of course, this might just tell the experimenter that he needs to improve his n(t) - build more temples.

[chrispedersen]

Originally Posted by analytic_kernel

Chris, I hope I'm not being too anal[ytical] about this, but I think that r is r(t), and that we don't necessarily know the form of n(t) or r(t). Looking at your number of posts, I'm guessing that you are a seasoned player, so I trust your judgement. But, if we wanted to be semi-empirical about it, we could collect data (possibly with the --statfile option, though I haven't tried it) on number of provinces and dominion for each turn, and then fit the data sets as a function of t. This would give the forms for n(t) and r(t).

Of course, this might just tell the experimenter that he needs to improve his n(t) - build more temples.

That actually would be kind of interesting. But the statfile gives no information on the number of temples... But still for an empirical fit...it would be ok.

Ah. But you still would have to adjust for *map* size, too, and density of starting positions...

[Sombre]

Originally Posted by analytic_kernel

Looking at your number of posts, I'm guessing that you are a seasoned player, so I trust your judgement.

This is not wise.

Edit: To be clear I'm not referring to Chris, just generally.

[MaxWilson]

Originally Posted by Sombre

This is not wise.

Heh. Without referring to chris in particular, I agree. I have a large number of posts and relatively little experience, none of it MP--people with many posts are *garrulous* players.

I tend to think the number of "thanks" are a more useful guide--at least you know that the person has had useful information to share in the past. Although there's still no guarantee that the really skilled players will show up on that metric, either.

See? I did it again--couldn't resist posting.

-Max