Tweet
Chaunk
This is a bit of an ongoing investigation, so I'll just post some initial findings.
1. SotB's free tech is free for one round of teching, but not after that: As the PK's I had the following tech costs:
No. Techs:
1 , 2,NA, 4, 5.
15,14,31,31,75
At 4 techs, the tech cost should be somewhere around the 45 mark for a large map. However, due to the free tech, it was 31, as expected for three techs. After researching the next tech, however, the cost jumped back to where it is expected to be for 5 techs discovered, around 70ish.
2. The uni get a research cost bonus. It's equivilent to a cost of n/(n+1), but it doesn't kick in until they've researched two techs. Their free tech is entirely free, and never counts towards their tech cost total.
3. The believers get a tech hit, equivilent to (n+1)/n cost. It does kick in at the first tech.
4. There seems to be a constant base cost for one tech, plus a penalty for doing well (and presumably a negetive for doing badly, but my testing hasn't proved that conclusivly yet - it's hinted at though as with the PK tech costs above going down after researching a tech). In a scenario situation, I added techs to the PK's and no other faction. Their tech cost started at 15, then jumped by 15+(4n) [very occasionally this formula breaks down due to the rounding effect. I suspect that 15 should be closer to say 15.55, and therefore occasionally an additional one is added to the cost.] every time until they had 17 techs, a tech lead of 15 techs. It then rose still, but by a smaller amount - the expected raise was 82, where the actual was 67. It was then increasing by ~2 additional points each time, not 4, i.e. 70 (Jump of three note... grumble!) points more, then 72 points etc.
Continuing the test past here, at 29 techs researched, the tech rise dropped again, this time to a constant 62 point increase every time.
The constant (around 15 here) is what I'd like to find out, as well as what affected the bonuses/penalties. It's interesting that I suspect that the tech cost itself never increases, but the player is merely penalised for doing well. In other words, the lowest teched person may go through the entire game with no increases above 15.
There's certainly more to come. I think the research formula may be slightly more simple than I'd imagined...
Chaunk
Additional point.
It seems likely that your first tech cost is the basic additional value, although I can't confirm that yet. That value is also rounded, seemingly rounddown, which can seem to spoil the standard running. Also, being somewhere other than first to research that first tech seems to almost always give you a slightly cheaper second tech - one point, or perhaps two points if you're lucky.
Impaler
Here is the most excrutiatingly complex tec cost Equation I have been able to find at Apolyton, it seems to work quite well though may be some inconsistencies that havent been found yet (though I belive you hinted at one of the major ones). With this equation you can predict tec costs FAR into the future and if combined with a prediction of your output you can nail down adsactly when the tec should be completed.
Cost = [ 11+ (DIFFICULTY x 4) + ((TEC x 6)/5) - (TURN/8) - (LEADER/5) ] x TEC x WORLD SIZE
DIFFICULTY = 1 to 5 with 1 being lowest and 5 being Trancendi
TEC = Number of tecnologies you currently have (or atleast how many are counted against you, as you noted their are exceptions)
LEADER = Number of tecnologies the faction that is tec leader has
TURN = turn of the game not counting 2100 so simply 1,2,3,4 ect
WORLD SIZE = modifier based on the worlds size, I belive its the square root of the maps tiles or something like that, once we know the size of the game map we can determine an adsact value for this and use it for any future evaluations.
-- Fear is the Mind Killer
AndiD
I may be or may be not related to this (or true after all):
(from the "Difference between supercollider and theory of everything SPs" thread at Apolyton)
Googlie
I didn't post in reply to S_C's 'Poly post 'cos i wasn't 100% certain ..... but ....
I believe that the lab points don't carry over to the next tech only from a base that singly is responsible for 50% or more of the research effort. I can't find the reference in either the official Guide or the Prima Guide, but i admit I haven't trolled 'Poly's threads to see if it's there. I think someone - MoSe (MariOne) did an analysis some years ago.
G.
AndiD
After an in-depth look I must state that the formula above doesn't give me any clue about tech costs, at least not on a Large map and on Transcend level.
The elements of that formula might be right (tech cost dropping because someone else researched tech or simply the year advanced, ...) but to find out how exactly each factor comes into play - not counting random-looking elements (penalty when mod3 kicks in for more than x techs? ...) - could well be just too much work for _one_ game.
I guess a random large map with each faction taking part in the ACDG, no AI control, tech addition by the scenario editor, manual turn advancement and energy generation by crawlers on energy specials (to get the lab points) will do the same trick.
For planning ahead, it is "only" important to know when a jump in cost will occur and in what ballpark that jump will probably be (= to estimate how we have to jump with lab point generation accordingly).
And of course, with every infiltration this research scenario would get more accurate (no. of techs & lab points & SE settings of our opponents).
Chaunk
For some reason I only just saw the replies to this thread. Impaler forumula
Cost = [ 11+ (DIFFICULTY x 4) + ((TEC x 6)/5) - (TURN/8) - (LEADER/5) ] x TEC x WORLD SIZE
seems good but somewhat flawed. I don't think there's any modification for the number of turns played at all. I'll post more evidence when I have it.
It's fairly easy to spot the pattern when it starts going, and I can now predict tech costs to within 20 points more or less all the time after the first few techs. Which surprisingly, is not a useful thing to be able to do at all really!
This is a bit of an ongoing investigation, so I'll just post some initial findings.
1. SotB's free tech is free for one round of teching, but not after that: As the PK's I had the following tech costs:
No. Techs:
1 , 2,NA, 4, 5.
15,14,31,31,75
At 4 techs, the tech cost should be somewhere around the 45 mark for a large map. However, due to the free tech, it was 31, as expected for three techs. After researching the next tech, however, the cost jumped back to where it is expected to be for 5 techs discovered, around 70ish.
2. The uni get a research cost bonus. It's equivilent to a cost of n/(n+1), but it doesn't kick in until they've researched two techs. Their free tech is entirely free, and never counts towards their tech cost total.
3. The believers get a tech hit, equivilent to (n+1)/n cost. It does kick in at the first tech.
4. There seems to be a constant base cost for one tech, plus a penalty for doing well (and presumably a negetive for doing badly, but my testing hasn't proved that conclusivly yet - it's hinted at though as with the PK tech costs above going down after researching a tech). In a scenario situation, I added techs to the PK's and no other faction. Their tech cost started at 15, then jumped by 15+(4n) [very occasionally this formula breaks down due to the rounding effect. I suspect that 15 should be closer to say 15.55, and therefore occasionally an additional one is added to the cost.] every time until they had 17 techs, a tech lead of 15 techs. It then rose still, but by a smaller amount - the expected raise was 82, where the actual was 67. It was then increasing by ~2 additional points each time, not 4, i.e. 70 (Jump of three note... grumble!) points more, then 72 points etc.
Continuing the test past here, at 29 techs researched, the tech rise dropped again, this time to a constant 62 point increase every time.
The constant (around 15 here) is what I'd like to find out, as well as what affected the bonuses/penalties. It's interesting that I suspect that the tech cost itself never increases, but the player is merely penalised for doing well. In other words, the lowest teched person may go through the entire game with no increases above 15.
There's certainly more to come. I think the research formula may be slightly more simple than I'd imagined...
Chaunk
Additional point.
It seems likely that your first tech cost is the basic additional value, although I can't confirm that yet. That value is also rounded, seemingly rounddown, which can seem to spoil the standard running. Also, being somewhere other than first to research that first tech seems to almost always give you a slightly cheaper second tech - one point, or perhaps two points if you're lucky.
Impaler
Here is the most excrutiatingly complex tec cost Equation I have been able to find at Apolyton, it seems to work quite well though may be some inconsistencies that havent been found yet (though I belive you hinted at one of the major ones). With this equation you can predict tec costs FAR into the future and if combined with a prediction of your output you can nail down adsactly when the tec should be completed.
Cost = [ 11+ (DIFFICULTY x 4) + ((TEC x 6)/5) - (TURN/8) - (LEADER/5) ] x TEC x WORLD SIZE
DIFFICULTY = 1 to 5 with 1 being lowest and 5 being Trancendi
TEC = Number of tecnologies you currently have (or atleast how many are counted against you, as you noted their are exceptions)
LEADER = Number of tecnologies the faction that is tec leader has
TURN = turn of the game not counting 2100 so simply 1,2,3,4 ect
WORLD SIZE = modifier based on the worlds size, I belive its the square root of the maps tiles or something like that, once we know the size of the game map we can determine an adsact value for this and use it for any future evaluations.
-- Fear is the Mind Killer
AndiD
I may be or may be not related to this (or true after all):
Santiago_Clause, 07.06.04 08:01:00
There is a reason not to colocate them: a base's research contribution is truncated when a discovery is made.
If you want to "tech sprint" by cranking up your science allocation you're much better off splitting SC and ToE and bringing their research output up evenly.
If all you really want is 1 tech per turn then colocate SC and ToE, set your labs allocation to whatever puts you just past your tech cost, and pocket the rest as ec's.
There is a reason not to colocate them: a base's research contribution is truncated when a discovery is made.
If you want to "tech sprint" by cranking up your science allocation you're much better off splitting SC and ToE and bringing their research output up evenly.
If all you really want is 1 tech per turn then colocate SC and ToE, set your labs allocation to whatever puts you just past your tech cost, and pocket the rest as ec's.
Googlie
I didn't post in reply to S_C's 'Poly post 'cos i wasn't 100% certain ..... but ....
I believe that the lab points don't carry over to the next tech only from a base that singly is responsible for 50% or more of the research effort. I can't find the reference in either the official Guide or the Prima Guide, but i admit I haven't trolled 'Poly's threads to see if it's there. I think someone - MoSe (MariOne) did an analysis some years ago.
G.
AndiD
After an in-depth look I must state that the formula above doesn't give me any clue about tech costs, at least not on a Large map and on Transcend level.
The elements of that formula might be right (tech cost dropping because someone else researched tech or simply the year advanced, ...) but to find out how exactly each factor comes into play - not counting random-looking elements (penalty when mod3 kicks in for more than x techs? ...) - could well be just too much work for _one_ game.
I guess a random large map with each faction taking part in the ACDG, no AI control, tech addition by the scenario editor, manual turn advancement and energy generation by crawlers on energy specials (to get the lab points) will do the same trick.
For planning ahead, it is "only" important to know when a jump in cost will occur and in what ballpark that jump will probably be (= to estimate how we have to jump with lab point generation accordingly).
And of course, with every infiltration this research scenario would get more accurate (no. of techs & lab points & SE settings of our opponents).
Chaunk
For some reason I only just saw the replies to this thread. Impaler forumula
Cost = [ 11+ (DIFFICULTY x 4) + ((TEC x 6)/5) - (TURN/8) - (LEADER/5) ] x TEC x WORLD SIZE
seems good but somewhat flawed. I don't think there's any modification for the number of turns played at all. I'll post more evidence when I have it.
It's fairly easy to spot the pattern when it starts going, and I can now predict tech costs to within 20 points more or less all the time after the first few techs. Which surprisingly, is not a useful thing to be able to do at all really!