Level;Cost;Efficiency;
0;1337;0;Base level
1;1419;50;
2;1500;60;
3;1550;65;
4;1600;70;
5;1650;75;
6;1700;80;
7;1750;85;
8;1800;90;
9;1875;95;
10;9999;100;Supa powa
4.3- ATM: Ahead of Time multiplier
This is a multiplier to penalize research when researching a tech ahead of the Ahead of Time Date (ATD).
ATD = (Start Year) - 5
If current date is ATD (see table above) or later, then ATM = 1
Else ATM = 1.5 + (ATD - Current Date) x 0.1
Examples:
Researching Land 20 in 1595. The ATD for Land 20 is 1595, so the multiplier is 1, as it will also be if the current date is later than 1595.
Researching Land 20 in 1593. The multiplier is 1.5 + (1595 – 1593) x 0.1 = 1.7
Chas said:The formula I use at the moment is:
Price = 0.22831 (End Year – 853.55), rounded to the nearest 0.5
This is a simple linear function. Note that the formula works for the beta, with just one of the constants changed:
Price = 0.22831 (End Year – 798.8)
Chas said:These produce all the right values for the dates in the actual .csv files, plus other dates between 1419 and 1819 that I have tested (which is not all of them).
It’s clear that they won’t really use these numbers in the program, the question is, what do they use?
Some things are definite:
- It really does only depend on End Date
- It is calculated by the program, not looked up – you can get values for any date, essentially.
- It really is linear. With judicious editing of files, you can get values from before 7000BC to after 9000AD. This gives fantastic accuracy on the estimation of linearity and the slope of the line.
- The price is definitely rounded (though exactly how is the big question) to 0.5.
Some things are nearly definite
- The slope is the same in the beta. At least, it’s within 4 decimal places the same, anyway.
The problem is that the formula above does not reliably replicate the program for all dates outside the ones I’ve mentioned. Mostly it does, but occasionally they are 0.5 out.
For a long time I thought the formula was:
Price = 300/70/1.02 (End Year – 855)
Beta = 300/70/1.02 (End Year – 800)
The idea was that perhaps in EU1, with a 300 year span, initially the span was 70d. Subsequently, when they change Catholic from 0 to +0.02 (i.e. a Tech Speed of -1) they adjusted the slope.
But it doesn’t quite work. So then I decided the slope was 823/400/9. The idea being that at some point they had divided by 9 to increase the values. This doesn’t qite work either.
In both these cases it is only a few values out of hundreds that round wrongly, but they do.
The formula might be expressed Price = Slope x End Year – Constant
Such minor changes affect possible rounding schemes. I’m pretty sure they are using some sensible integer values, in some combination, with some mix of integer and floating point arithmetic to do the calculation and provide the rounding.
But I haven’t cracked it, so I haven’t put the formula in the FAQ – it does produce all the BV’s in the FAQ, but I know it isn’t quite right.
Anybody that solves it will earn my eternal admiration and gratitude.
No, the ATD for Infra 5 is 1550-5 = 1545.Yes, I read that. But that doesn't explain the situation I gave. Basically people researched techs way before respective ATD.
For example, from the stats I posted, Venice finished infra 5 by 1560.
The ATD for infra 5 is 1645 (1650-5)
ATM = 1.5 + (1645-1560)*0.1 = 10
Could Venice have possibly researched infra 5 with an ATM of 10 and more?
Chas said:Code:0 1419 1 1425 Gives the End Year for level 2, i.e. we are at level 1 2 1450 3 1495
Start year for level 5 is red. End year is yellow. ATM = Start year - 5.Level;Cost;Efficiency;
0;1337;0;Base level
1;1419;50;
2;1500;60;
3;1550;65;
4;1600;70;
5;1650;75;
6;1700;80;
7;1750;85;
8;1800;90;
9;1875;95;
10;9999;100;Supa powa
No, the ATD for Infra 5 is 1550-5 = 1545.
Well, they would get Ahead of Time penalty if they are too far ahead (as determined by the .csv files). It's not detrimental per se, as they are ahead in tech. Also, they are unlikely to put so much research into a tech that they manage to go over the usual cost before the time has catched up to them. Ie, if the cost is 1000d with ATM = 1, and with a higher ATM it's 3000d, then you will only "lose" if you research more than 1000d in that tech.Is it detrimental for states like VEN, GEN, TUR etc etc to start the GC at levels higher than the correct figure of 1 ??
Nothing. The ATM is 1, so the Ahead of Time penalty is nil. There is no "After Time bonus" (except for neighbour bonus).Btw, what happens if you research a tech after the end year?