Tweet
Originally posted by PJayTycy
what's the cost for distance 0 (ie: mind control the headquarters itself) ?
what's the cost for distance 0 (ie: mind control the headquarters itself) ?
-
Probes can't mind control an opponent's HQ."The avalanche has already started. It is too late for the pebbles to vote."
-- Kosh -
This data is easy to fit if you look at it a bit. I took the listed mind control costs and plotted their inverses against distance. It's nearly a perfect line, and their correlation is .99995, nearly perfect. I calculated linear regression parameters and ran the distance values back through the calculated line. The resulting costs were pretty good, but were off by an EC in places, due to rounding present in the original data. I looked a little harder and realized that the intercept was about 4 times the slope, ran some calculations with the listed distance values plus 4, and got a slope of about 1/1200.
I'd seen the number 1200 before. It was the number of ECs my opponent needed to double my mind control cost in some crude tests I'd done. Based on all this, I present a somewhat untested mind control formula:
d = distance from HQ (after modification by creche or genejack)
E = opponent's ECs
p = population of opponent's base
C = other stuff potentially including number of bases already mind controlled, opponent's probe rating, etc. By default, 1
Cost to control(d, E) = C * p * (1200 + E) / (d + 4)
After rounding, the results match up perfectly for size 1 bases owned by opponents with 10 ECs. The cost to mind control such an opponent's HQ, if it were possible, is 302, by the way."Cutlery confused Stalin"
-BBC newsComment
-
Children's creche's halve the distance, and genejack factories double the distance, however, I don't know if this includes just halving the distance or halving the distance plus 4...
By the way, what do you mean when you say inverse?Comment
-
Note how the distance is in the denominator? To derive this, I took all the energy values posted by bluetemplar and used their reciprocals in calculating the line of regression. I could have tried the reciprocals of the distances, but due to the +4, that wouldn't have worked."Cutlery confused Stalin"
-BBC newsComment
-
I think I understand, as I know what reciprocals are...
But how would children creche's fill in to this equation? I know that if population is one and the AI has no energy, the cost is effectively the same as if the distance is halved if a children's creche is present...I dont' think the variable C represents this...Comment
-
According to Chaos, CC are included in d :
* d = distance from HQ (after modification by creche or genejack)
With or without religion, you would have good people doing good things and evil people doing evil things. But for good people to do evil things, that takes religion.
Steven WeinbergComment
-
Sorry...
I bow down to the wisdom of Chaos Theory...
Comment
-
Needs some testing. Comments coming soon.
New version:
1.01 : Unit Row Cost Calculator added
Another version:
v1.02
Base Cost : depends on number of bases captured this turn
Probe succes rate table addedLast edited by bluetemplar; June 9, 2005, 14:57.Comment
-
-> Sometimes for the distance calculator to work, you must invert the positons of unit/base and HQ.
I fear that the next three weeks I will have no time to spend to SMACX and to clear all the bugs that are in this file...
what i don't understand is why the Unit Row Cost Calculator doesn't work while it's the exact formula that is written in the game help!
Damn, I should have thought about an hyperbola, since it wasn't an exponential, nor parabola, neither it couldn't have been a polynomial. But the polynomial was closer : about 0.9996 against 0.9990 for a reciprocal.
There is a link for the software I used for the other regression (units) for my calculator (it was easy now since i knew what sort of function it was, thanks Chaos Theory )
But I still hadn't found what is the formula for genejack, it's near distance*3 but not exactly...
Comment
-
Originally posted by bluetemplar
what i don't understand is why the Unit Row Cost Calculator doesn't work while it's the exact formula that is written in the game help!
You mean the eight-term polynomial was closer? Sure, since it has eight degrees of freedom to a hyberbola's two. Your polynomial would have quickly diverged to positive or negative infinity outside the range of data fitted, though.Damn, I should have thought about an hyperbola, since it wasn't an exponential, nor parabola, neither it couldn't have been a polynomial. But the polynomial was closer : about 0.9996 against 0.9990 for a reciprocal."Cutlery confused Stalin"
-BBC newsComment
-
I can't wait till civ4, no more having to figure out equations...Comment
Comment