100 / 2 > 50 ?

Hello.

I have a question about Civ4 and Micro-management. Please forgive my poor English. I'm better at translating poetry than maths.

Anyway, let's suppose your city produces 6 commerce and has a library. Suppose maintenance and other costs limit your research capacity to, say, 50% of your income. Then, aren't you better off investing 100% half of the time and 0% the other half, instead of permanently investing 50%? If you stick to 50%, you probably only get 3 Research Points per turn, due to numbers being rounded down (not sure about this). But if you follow my method, you alternately get 7 and 0 RP, resulting in an average 3.5.

Has anyone experience to share or do you know where I can find reliable information about this?

Yes, there is an argument for running at 100% and 0% science and absolutely nothing in between, since it will give you very slightly more research by eliminating rounding errors as Civ always rounds fractions down. I don't bother with it myself though, since it can generate an absolute maximum of 1 beaker a turn, and over the course of the game this really doesn't add up to that much, while it wastes a lot of of time messing around with the slider. What's say 400 beakers over the course of a normal game? That's what, 5% of a modern tech? That's the most you can get, and in practice it will be lower as the rounding will frequently lose less than the full 0.9 beakers a turn. It doesn't even get you the earlier techs significantly earlier as the gain is so finely spread. It also doesn't work if you're using the culture slider for happiness, since you then can't run at 0% or 100% science without creating unhappiness.

If you really like to micromanage there is a tiny gain to be had here, but it's a lot of effort for very little return.

EDIT: I thought I must have forgotten something here . Yes, it's per city, so assuming a decent size empire you might get most of a modern tech over the course of a game. It does become more time consuming at the start at the higher levels though as you can't run 100% straight out the door.

Actually, it can save one beaker per city per turn, which really amounts to a lot. Also, there's actually very little micromanagement involved when doing it at the lowest level.

MrCynical said:

It doesn't even get you the earlier techs significantly earlier as the gain is so finely spread. It also doesn't work if you're using the culture slider for happiness, since you then can't run at 0% or 100% science without creating unhappiness.

Click to expand...

The effect is greater in the earlier years, so it does get you ancient techs earlier. The culture slider doesn't negatively impact the strategy at all, since you can apply the calculations to whatever is left after culture just as well as you can apply it to the whole 100%. The only difference is that it can't then be coupled with the "commerce in multiples of 4" trick.


To read more about binary science rate, go there : http://forums.civfanatics.com/showthread.php?t=159109

In another thread, someone calculated that you can potentially save .7 beakers/turn/city. In a Normal game, the mean average cost of a tech is 7530 beakers (Future Tech, 15000 beakers + Hunting or Fishing, 60 beakers divided by 2). Supposing an empire size of 10 cities, you can save 3220 beakers during the course of the game (7 beakers/turn X 460). That's not even enough to gain one free, mean average tech. You'd need an empire of at least 25-30 cities to really make it pay off. That also doesn't consider the number of turns it takes to get an empire of that size.

Well, thanks everyone for the info. I think I'll continue doing this after all.
The most tedious part of it is of course clicking ten times in a row. I really wish you could just hold down the mouse button to achieve this.

Willem said:

In a Normal game, the mean average cost of a tech is 7530 beakers (Future Tech, 15000 beakers + Hunting or Fishing, 60 beakers divided by 2).

Click to expand...

Not true. The average cost of techs is the costs of all techs, divided by the total number of techs. It's much lower than 7530 because there are a lot more cheap techs than expensive ones. A geometric average, which might make more sense in this case, would give an even smaller amount. I would guess the geometric average cost of tech to be somewhere around 500 beakers.

Willem said:

Supposing an empire size of 10 cities

Click to expand...

That's a pretty small empire. Mines are usually around 20 cities.

Willem said:

You can save 3220 beakers during the course of the game (7 beakers/turn X 460).

Click to expand...

Or 6440 with a more standard 20 city empire. Enough for about 12 techs using the geometric average, or one single late game tech (and allowing you to get all techs earlier). Very significant.

Zombie69 said:

Not true. The average cost of techs is the costs of all techs, divided by the total number of techs.

Click to expand...

I did say mean average. I'm not about to add up all the tech costs in the game and find the true average. Feel free to do so if you like.

That's a pretty small empire. Mines are usually around 20 cities.

Click to expand...

I used 10 cities simply for convenience sake.

Very significant.

Click to expand...

Not for all the trouble it requires IMO. There's already enough micromanagement in the game.

Willem said:

I did say mean average. I'm not about to add up all the tech costs in the game and find the true average. Feel free to do so if you like.

Click to expand...

Mean is synonymous with average. Saying "mean average", a useless repetition, doesn't change the fact that you didn't count it correctly.

mean, mean value
an average of n numbers computed by adding some function of the numbers and dividing by some function of n

Willem said:

I used 10 cities simply for convenience sake.

Click to expand...

And this had nothing to do with the fact that it helped your point? You've shown throughout the threads that you either have very little understanding of math, or very little regard for honesty.

Willem said:

Not for all the trouble it requires IMO. There's already enough micromanagement in the game.

Click to expand...

Requiring micromanagement doesn't make something insignificant. Something can be very significant and at the same time require significant effort. For you, the cost may not outweigh the reward. But that's not a reason to dish the reward and try to make it seem like it's not there.

With your obsession to try and disprove something that has already been proven by many, the only logical explanation i can find is that you're not actually trying to convince us that the effect is minimal. Rather, you're trying to convince yourself that it doesn't really matter so you can feel better, wrapped in the idea that you're not really missing out on a lot anyway and feeling less guilty that you're not sharing in this bonus because you can't be bothered to work to obtain it.

It's perfectly ok to not want to do binary science. But it's quite another thing to try to convince people that it's useless despite all evidence to the contrary. Do what you want in your games, but stop the misinformation.

The 10 cities would also be the average over the course of the game. There are a LOT of turns going by with less than 5 cities on a standard map. Still, the average city number might be 20, in a complete game and a big empire. It depends.
However, the bulk of the Beakers gained would be right near the end of the game, when techs are most expensive.
Zombie is correct in that a geometric average would be the most useful. But who's volunteering to do all that work?? lol!
Are we also factoring in that there will be turns of zero beakers? I think the origional conclusion, that there is some gain, but it's really not signifigent, is truest. After all, the AI could easily do that much micro-managing, if there were a big benifit, that's how it would play, eh?

The trouble? What trouble? What is this trouble you are referring to? Sliding from 0 to 100 is that it??? Give me a break
There are tons of advantages to running 0 and 100 ,
1 - Gets you more techs
2 - In case of emergency spend the money you have for upgrades
3 - The technology you were planning to get before ai (for example liberalism) has been discovered by another ai. Now there is another tech priority - not a problem, just switch to whatever you need. Not very convenient when you are in the midd of tech you are researching, is it?

Its just superior way of play, but if you cant be bothered to switch that slide now and then ok your thing. Personally i think its peanuts and don’t see “the trouble” or “effort” people are referring to... mostly the people who don’t yet understand the math or concept behind it…

5cats said:

Zombie is correct in that a geometric average would be the most useful. But who's volunteering to do all that work?? lol!

Click to expand...

It would take about 30 seconds to work out in Excel provided i had the table with all technology costs which was already posted in a thread somewhere, but i don't know which thread and don't feel like looking all over the place for it.

Copy-paste the table here and i'll give you the geometric average if you want. I suspect it's actually below 500, maybe as low as 300 but probably somewhere around 400. But this is just a guess.

5cats said:

I think the origional conclusion, that there is some gain, but it's really not signifigent, is truest. After all, the AI could easily do that much micro-managing, if there were a big benifit, that's how it would play, eh?

Click to expand...

No, because the AI is stupid and will only do what it was programmed to do. Since binary science is probably not the way the designers had in mind of using the science slider, they didn't program the AI to us it this way.

A true AI, however, would surely use binary science all the time!

No doubt about it.

Zombie69 said:

It would take about 30 seconds to work out in Excel provided i had the table with all technology costs

No, because the AI is stupid and will only do what it was programmed to do. Since binary science is probably not the way the designers had in mind of using the science slider, they didn't program the AI to us it this way.

A true AI, however, would surely use binary science all the time!

Click to expand...

Aye, there's the rub! It's entering all that data that's a pain in the A, of course it'd be easy from there. I might go and look for it.

True! But only if the AI had freewill

Really, we'd need all the data from many games to accurately figure this out. Without knowing exactly how many Beaker fractions are lost, and when, it's just speculation. Of course if someone does know those numbers, or how to find them, clue me in!

Zombie69 said:

A geometric average, which might make more sense in this case...

Click to expand...

I'm curious, why might you use the geometric mean for tech costs? I know the definitions but am not familiar with the science of choosing between them. I'm a statistician by training, and the only practical application I'm aware of is that it makes more sense to compare arithmetic means when data is normally distributed, but geometric means when data is lognormally distributed. A google search led me to several financial arguments that I don't understand, and I'm suspecting your reasons might be related to them.

FYI:
arithmetic mean = (x1 plus x2 plus ... plus xn) / n
geometric mean = (x1 times x2 times ... times xn) ^ (1/n)

Zombie69 said:

It's perfectly ok to not want to do binary science. But it's quite another thing to try to convince people that it's useless despite all evidence to the contrary. Do what you want in your games, but stop the misinformation.

Click to expand...

All I'm trying to do is offer the other side of the story. You're making it out to be something you need to do in order to do well in the game, like it's some fabulous strategy that you'd be a fool not to pursue. Which simply isn't true. It's only misinformation if people see just one aspect of the strategy, which you seem compelled to do by trying to intimidate anyone one who contradicts your views.

Working on the costs at Noble level and Normal speed, I find the total cost of 83 techs to be 242,346 at an average cost of 2920. There are 31 techs more expensive than that average, and 54 cheaper (of which two are given as starters, valued at 52, 65 or 78 each). It is unreasonable to include Future Techs, since the number of these is unlimited.

While it is true that running at 100% makes no waste, this is not always possible. Your economy may not support such a rate, or you may need to spend something on Culture to keep your citizens happy or work towards city expansion. And the amount of waste whilst researching a particular tech depends heavily on how many turns that research is going to take: the waste per turn can never exceed 0.9 points (and then only if you're running at 90%) which is fairly trivial if the research will take two or three turns but does start to matter if it'll be 20 turns or so. Anyway, the total waste over 460 turns will not be more than 460*0.9 = 414 flasks, out of a total of over 240,000 if you do actually develop all techs other than Future and haven't won before time runs out.

Edit: yes, it's per city. So while you have few cities, say up to 500BC, the possible waste won't really matter. On another aspect I'd hate to try working out the geometric mean as I reckon the figure obtained by multiplying together all the separate tech costs would probably burst my machine: at a very rough estimate it would be in the region of 3*10^257.

morchuflex said:

Well, thanks everyone for the info. I think I'll continue doing this after all.
The most tedious part of it is of course clicking ten times in a row. I really wish you could just hold down the mouse button to achieve this.

Click to expand...

I hear ya! This is on my list for the mod I want to make, as soon as I can buckle down and get to it... separate buttons for "0% science" and "100% science." Or maybe even simplifying things and encouraging better performance by re-mapping the existing buttons to that . I'm also going to add a toggle to make actions last only one turn so that you don't have to go back to each unit to hit the freakin' cancel button each time! Zombie69, you seem to do a lot of micro-management with workers-- do you know of any mods that do this already?

Bushface said:

Anyway, the total waste over 460 turns will not be more than 460*0.9 = 414 flasks, out of a total of over 240,000 if you do actually develop all techs other than Future and haven't won before time runs out.

Click to expand...

Total per city will not be more than that - if you've an average, say, 15 cities over the course of the game (fewer in the beginning, more at the end) that adds up to a maximum possible waste of 6212 flasks, about 2 techs, and that of course is assuming that your average 15 cities each produce an odd number of commerce points every single turn and that you're not trying to make any money or use the culture slider, and that you're comparing your MMing to someone running 90% science on average (unfeasibly high except on an OCC, where of course you don't mind MMing anyway, but your maximum lost research through not doing binary research is 414 flasks over the game anyway) which would be the most wasteful.

At worst, not doing binary research will cost you maybe 1 tech's worth of research compared with someone who is doing it, and the endless micromanagement and keeping track of how many turns you need to run 100% for (and how many at 0%) would drive you insane before you got to finish it.

Better: 80% science most of the time. Build courthouses, markets, banks, and grocers to increase income. Keep the people happy. The result is We Love the King Day -- no maintenance costs.

Also, 50% now for 10 turns, then 50% later for 10 works out if you build an economic upgrade. But how are you going to schedule all your cities to get their banks built at the same time?

I suggest you leave it alone.

So lets see if I got it straight:
The maximum loss per city = 414
The average loss per city = 207
Then factor in the number of turns you're going to run at 0% science, because otherwise you'll go bankrupt.
So it works out to being less than 1 extra tech for the entire game's duration.
Keep in mind that the more cities you have, the more gold you're going to need, which means less turns at 100% science. So while having a large empire & MMing it would save more beakers, this would be countered by having more turns at 0% science to pay for all those cities.

Conclusion: All that Micro Managing will produce less than 1 technologly in an average game. This means that insted of learning 50 techs (and trading for others & etc) you'll learn 51 by 2050.