Cromagnus
Deity
- Joined
- Sep 11, 2012
- Messages
- 2,272
Well, I had a chunk of time this weekend, and made it to t438.
I'm now faced with a dilemma. Does anyone know at what point during the calculations the fixed point limit of 214k beakers is applied? IE is it an intermediate value that is limited to 214k or is the final value left over?
This is a real problem because I'm at 37k bpt. If I bulb, I get 296k beakers, which is over the limit. But let's assume the intermediate value is preserved as a 32-bit number.
Techs are 107k. If I currently have zero overflow, my overflow will be 296k-107k=189k. (Possibly 185k because overflow is now limited to 5x bpt)
If I then hit next turn, I'll hit 226k, which is over the limit. But if the tech cost is applied before the limit, I'll be at 119k, and everything is fine.
The only case I'm worried about is the 2 possible intermediate values being stored in fixed-point.
If 296k is stored in fixed point, it will be translated as a negative number, and I'll *lose* turns of beakers by bulbing. If the 226k is stored in fixed point, it's an even larger negative number.
However, if overflow is only stored as fixed point after techs are applied, then I think I'm fine, because in both of the above cases, the subtraction would prevent the overflow.
Anyone know for sure what will happen? I'm already getting 1 Future Tech every 3 turns, so I might just plant the stupid GS as an academy and not risk it. But I'm curious.
Also, a few key updates for those who may attempt this later:
1) At t438 I've already exceeded my prediction of 15k score, but population growth is diminishing fast. I'll probably finish with about 17k, which is probably about 2k below the potential score with a better map and an earlier WF. So for those trying to get an unassailable gold for the HOF, you'll probably need to hit 19k+.
2) I'm at ~500 Happiness. My obsession with getting the ice&snow-locked cities to the happiness cap was *totally* misguided. I see that now. Once a city hits size 14 it has enough population that your local happiness isn't capped. (max local happiness = population)
Because of the reduced unhappiness from specialists of roughly 5.5/city, reduced unhappiness from liberty and Forbidden Palace, and religious + liberty global happiness, a city of size 14 with full buildings generates like +7 happiness, which can be used to compensate for cities that exceed the happiness they generate. I was capping my fastest-growing cities at 32 to leave happiness room for the other cities to grow into, but because not all cities will have time to cap out, and some don't have the food, there's no point in doing so.
This leaves you free to send trade routes to the cities that *benefit the most*, IE cargo ship always, never caravans. (Except possibly early on to get a city to size 14)
I say 14 but 21 is better. 14 * 0.85 = 12, - 1/10 = 11. 21*0.85 = 18, -2/10 = 16. The population unhappiness reduction on a size 21 city is noticeably better than a size 14. But it can be difficult to get a city to size 21 when they have zero workable food tiles...
Coliseum + Zoo + Stadium + Pagoda + Temple + Bank + Stock Exchange + Hospital + Medical Lab + Garrison = 15 local happiness. So 14 is almost enough in the worst case. But in the typical case, your snow-locked city is at the bottom of the map, so technically it's coastal, and gets an extra +3 for being coastal. So you need at least 17 population to get the max happiness value of that city. Possibly 21. if the city has Stone Works/Mint/Water Mill. Which is another reason 21 is better. But at least 17.
Point being... sure, it's not as efficient, and the cities at size 32 grow slowly, but a one-tile island has 36 food tiles around it, and with King Days and everything else, even at size 47 it will still have a very significant food surplus. (ToA really adds up when your base food is 90...)
The only thing that matters is that every city generate the max local happiness it can.
3) One last thing about happiness. +500 seems like a lot, sure, but with 180 cities, that's not even +3/city. Meaning that I could conceivably go negative before the end of the game. When I was at +220 I thought it was wasteful, but with 50 cities growing each turn, even -1/city would have taken me negative temporarily. (Like, without garrison happiness for example)
So, perhaps I was wrong about the happiness policies being excessive...
Anyway, now I'm also faced with the possibility that my total gold could go negative. Hopefully that isn't capped at 214k also. Even building research, I'm at 13k gpt, and I have nothing to buy except nuclear/solar plants. (To eke out a few extra beakers)
I may just rush-buy air units in every city to keep that number under 214k. It'll make capturing the last CS hilarious at least.
I'm now faced with a dilemma. Does anyone know at what point during the calculations the fixed point limit of 214k beakers is applied? IE is it an intermediate value that is limited to 214k or is the final value left over?
This is a real problem because I'm at 37k bpt. If I bulb, I get 296k beakers, which is over the limit. But let's assume the intermediate value is preserved as a 32-bit number.
Techs are 107k. If I currently have zero overflow, my overflow will be 296k-107k=189k. (Possibly 185k because overflow is now limited to 5x bpt)
If I then hit next turn, I'll hit 226k, which is over the limit. But if the tech cost is applied before the limit, I'll be at 119k, and everything is fine.
The only case I'm worried about is the 2 possible intermediate values being stored in fixed-point.
If 296k is stored in fixed point, it will be translated as a negative number, and I'll *lose* turns of beakers by bulbing. If the 226k is stored in fixed point, it's an even larger negative number.
However, if overflow is only stored as fixed point after techs are applied, then I think I'm fine, because in both of the above cases, the subtraction would prevent the overflow.
Anyone know for sure what will happen? I'm already getting 1 Future Tech every 3 turns, so I might just plant the stupid GS as an academy and not risk it. But I'm curious.
Also, a few key updates for those who may attempt this later:
1) At t438 I've already exceeded my prediction of 15k score, but population growth is diminishing fast. I'll probably finish with about 17k, which is probably about 2k below the potential score with a better map and an earlier WF. So for those trying to get an unassailable gold for the HOF, you'll probably need to hit 19k+.
2) I'm at ~500 Happiness. My obsession with getting the ice&snow-locked cities to the happiness cap was *totally* misguided. I see that now. Once a city hits size 14 it has enough population that your local happiness isn't capped. (max local happiness = population)
Because of the reduced unhappiness from specialists of roughly 5.5/city, reduced unhappiness from liberty and Forbidden Palace, and religious + liberty global happiness, a city of size 14 with full buildings generates like +7 happiness, which can be used to compensate for cities that exceed the happiness they generate. I was capping my fastest-growing cities at 32 to leave happiness room for the other cities to grow into, but because not all cities will have time to cap out, and some don't have the food, there's no point in doing so.
This leaves you free to send trade routes to the cities that *benefit the most*, IE cargo ship always, never caravans. (Except possibly early on to get a city to size 14)
I say 14 but 21 is better. 14 * 0.85 = 12, - 1/10 = 11. 21*0.85 = 18, -2/10 = 16. The population unhappiness reduction on a size 21 city is noticeably better than a size 14. But it can be difficult to get a city to size 21 when they have zero workable food tiles...
Coliseum + Zoo + Stadium + Pagoda + Temple + Bank + Stock Exchange + Hospital + Medical Lab + Garrison = 15 local happiness. So 14 is almost enough in the worst case. But in the typical case, your snow-locked city is at the bottom of the map, so technically it's coastal, and gets an extra +3 for being coastal. So you need at least 17 population to get the max happiness value of that city. Possibly 21. if the city has Stone Works/Mint/Water Mill. Which is another reason 21 is better. But at least 17.
Point being... sure, it's not as efficient, and the cities at size 32 grow slowly, but a one-tile island has 36 food tiles around it, and with King Days and everything else, even at size 47 it will still have a very significant food surplus. (ToA really adds up when your base food is 90...)
The only thing that matters is that every city generate the max local happiness it can.
3) One last thing about happiness. +500 seems like a lot, sure, but with 180 cities, that's not even +3/city. Meaning that I could conceivably go negative before the end of the game. When I was at +220 I thought it was wasteful, but with 50 cities growing each turn, even -1/city would have taken me negative temporarily. (Like, without garrison happiness for example)
So, perhaps I was wrong about the happiness policies being excessive...
Anyway, now I'm also faced with the possibility that my total gold could go negative. Hopefully that isn't capped at 214k also. Even building research, I'm at 13k gpt, and I have nothing to buy except nuclear/solar plants. (To eke out a few extra beakers)
I may just rush-buy air units in every city to keep that number under 214k. It'll make capturing the last CS hilarious at least.