Micromanagement is alive and well in Civ 4!
- Thread starter Zombie69
- Start date
ruff_hi
Live 4ever! Or die trying
I've been slowly working through the beaker to tech calculations and also getting feedback on my thoughts. To summaries:ruff_hi said:
- base commerce gets converted to base beakers based on the science slider percentage - done at a city level (rounded down)
- base beakers are adjusted for science multipliers (library, etc) - done at a city level (rounded down)
- total base beakers (all cities) are adjusted by other civs knowing tech and you knowing pre-requisites (rounded down)
- adjusted total base beakers are applied to research tech
As I understand it, binary science only works if you are running in the red @ 100%. If you aren't, then hang the lost beakers and keep reseaching. If you are running in the black @ 90%, but in the red @ 100%, then the binary science proposition is to run 1 turn @0% and 9 turns @ 100%. Thus you hit the 90% overall and you minimize beaker loss.
The balance of this post is going to look at this beaker loss.
Here is an example using various base commerce of the beakers that are lost via 2) assuming the city has a 75% science multiplier.
Detailed Calculations
Science Slider @ 90%
Base Commerce 11
Science Multipler 75%
Base Beakers = 11 * 90% rounded down = 9.90 rounded down = 9
Adj Base Beakers = 9 * (1.75) = 15.75
Lost Beakers = 15.75 - 15.00 = 0.75
Summarized Results
Base ............. Slider Percentage
Commerce ..... 70% ..... 80% ... 90% .... 100%
... 05 ......... 0.25 .... 0.00 .... 0.00 .... 0.75
... 06 ......... 0.00 .... 0.00 .... 0.75 .... 0.50
... 07 ......... 0.00 .... 0.75 .... 0.50 .... 0.25
... 08 ......... 0.75 .... 0.50 .... 0.25 .... 0.00
... 09 ......... 0.50 .... 0.25 .... 0.00 .... 0.75
... 10 ......... 0.25 .... 0.00 .... 0.75 .... 0.50
... 11 ......... 0.25 .... 0.00 .... 0.75 .... 0.25
... 12 ......... 0.00 .... 0.75 .... 0.50 .... 0.00
... 13 ......... 0.75 .... 0.50 .... 0.25 .... 0.75
... 14 ......... 0.75 .... 0.25 .... 0.00 .... 0.50
... 15 ......... 0.50 .... 0.00 .... 0.75 .... 0.25
... 16 ......... 0.25 .... 0.00 .... 0.50 .... 0.00
... 17 ......... 0.25 .... 0.75 .... 0.25 .... 0.75
... 18 ......... 0.00 .... 0.50 .... 0.00 .... 0.50
... 19 ......... 0.75 .... 0.25 .... 0.75 .... 0.25
... 20 ......... 0.50 .... 0.00 .... 0.50 .... 0.00
So, you can see from the above table that running at 100% has no wasted beakers when your city level commerce is a multiple of 4. Any other city level commerce wastes either 0.25, 0.50 or 0.75 beakers per city. However, if you look at running at 80% - it wastes 0.00 beakers when your city level commerce is a multiple of 4 AND when it is 1 less. Edit:Oops, should be multiple of 5 and multiple of 5 + 1
To me, this means that running binary science @0% and @80% is better than running it at 0% and @100% because there are additional city level base commerce amounts were you lose no (partial) beakers (again - assumes that you are losing money @ 80%). So, if you are breaking even @70%, then instead of running at 0% for 3 turns and 100% for 7 turns, run at 0% for 1 turn and 80% for 9 turns (equivalent to running at 72%).
I think an additional benefit is that you don't have to adjust your science rate as often and you have more extra gold longer.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
ruff_hi said:
This is only true when you have a single city, and no gold bonuses (market, bank, grocer, wall street). As soon as you have a few cities, binary science at 0% and 100% becomes optimal because different cities will have different base commerce and different commerce multipliers. The only way then to minimize lost fractions is to always run at 0% (thus losing no fractions on science bonuses since you make no science) or 100% (thus losing no fractions on gold since you make no gold). At any other percentage, you open yourself up to losing fractions both on beakers and on gold in each of your cities, instead of losing fractions on only one of those.
Also, note that by running science at 80%, you're running gold at 20%, which you've already proven is a bad proposition with a 75% bonus on gold, and i can already tell you is also a bad proposition at 25% or 50% bonus.
Remember, the point is not only to avoid wasting fractional beakers, but also to avoid wasting fractional gold. After all, 1 gold is just as important as 1 beaker (in fact it's even more important if your average science rate is above 50% for reasons which i won't get into).
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
ruff_hi said:
Wrong. From your table, it's clear that no fractional beakers are wasted at 5n and 5n+1, not at 4n and 4n-1. This makes a difference because it means no lost beakers 40% of the time (i.e. in 2 out of 5 cases), not 50% of the time (i.e. in 2 out of 4 cases).
Regardless, 80% science is no good because of the two aspects mentioned in my post above, namely :
1. Multiple cities with different % bonus and different base science.
2. Bonuses to gold being applied to 20%, a bad percentage.
Zombie69 said:
It is quite common for me to have a few specialized cities, where +90% of my research is done. When I get writing I build libraries in them ASAP. So if I have 5-6 cities with libraries (accompanied by 2-3 cities that only produce units and have negligable gold,) it may make sense to maintain 80% - 0%, correct?
When I get universities I build universities in them ASAP. I really stress building 6 universities ASAP, because I love Oxford university. Since I beeline for liberalism after writing, I often get universities prior to marketplaces. If I have 6 cities with universities, prior to Oxford I wonder if 100% - 50% might work just as well as 100% - 0%?
Anyway, this thread has definately intrigued me and I will use binary science next GOTM, at least until I get libraries.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
When i get writing, i bring my science rate down to 0%. I keep it there until most or all of my commerce cities have built their libraries. Then i can switch back to 100% for a very long time. I do the same thing with education and universities, usually waiting until i get not only the universities but also Oxford.
This allows you to make full use of your new buildings, for more turns than you would otherwise.
This allows you to make full use of your new buildings, for more turns than you would otherwise.
ruff_hi
Live 4ever! Or die trying
Yes, well spotted. I looked at 16 commerce and 15 commerce and thought 16/4 and (16-1)/4. I should have looked at the other combinations.Zombie69 said:
One tiny point - I know that you have put up with a lot of angst with this discussion and people getting their maths wrong - but I have been posting and slowly discussion this for a while, accepting corrections when they are pointed out. You must know that the written word comes out differently than the spoken work - much harsher. Thus your "Wrong." really does grate. Sure if you have pointed out things multiple times, but with me - that is not the case. It would have come across much better as "Your comment about 4n and 4n-1 is incorrect.".
I'll start thinking about the other side of the coin (so to speak) - gold. I read another post that floated the idea of science sliders, not at the civ level, but at the city level. The initial reaction was YES PLS. But on reflection the group decided that it would unbalance the game too much.
DaviddesJ
Deity
ruff_hi said:
If you're talking about how to fix the micromanagement, it's easy---just allow the +25% bonus from the library to create fractional beakers that get credited to research. Individual city sliders would go in the wrong direction---even more micromanagement.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
DaviddesJ said:
True. Adding fractions over all cities would drop the max amount of beakers saved by binary science to 1/turn as opposed to 1/city. And if you keep any fraction still left and add it to next turn's total, then you reduce the effectiveness even more, droping it to a maximum of 1 beaker over the entire game.
Of course, binary research would still be useful for the extra bonuses obtained from civs knowing the tech, the options provided by having extra money in the bank most of the time, and the extra flexibility afforded by being able to choose the tech researched a few turns later.
In the case of a lost fraction, could it be the optimum solution to use a specialist instead of a worker on a tile, in a "per city" micromanagement? Of course it depends on the food, but still the number 3 looks like it's made for "adjusting" the lost fractions due to the division by 4.Zombie69 said:
Like I said, I'm intrigued by binary science and I'm going to give it a try. Everyone plays differently, (I play for fastest finishes), but I can't see a couple of these binary science bonuses helping me.
This won't help me because I disagree with this kind of strategy. I never research a tech if another civ knows it. I typically start BW + pottery + writing + ABC. I beeline for alphabet and then trade like crazy. It is MUCH more efficient to never research a tech at all and to trade for it than to research it at a discount.
This is an interesting notion, and it may have its value, but I have never switched research in the middle of researching a tech in a serious (i.e. GOTM) game.
Zombie69 said:
This won't help me because I disagree with this kind of strategy. I never research a tech if another civ knows it. I typically start BW + pottery + writing + ABC. I beeline for alphabet and then trade like crazy. It is MUCH more efficient to never research a tech at all and to trade for it than to research it at a discount.
Zombie69 said:
This is an interesting notion, and it may have its value, but I have never switched research in the middle of researching a tech in a serious (i.e. GOTM) game.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
atreas said:
Doubtful, since in terms of production (i.e. ignoring GPP), specialists are rather pathetic. Still, sometimes i will do this. More often however, i will adjust the number of specialists to avoid lost fractions of GPP. For example, with a 50% bonus in GPP production, you want your base number of GPP per city to be an even number. Another thing i'll do in a science city is drop specialists and put them on small cottages to let them grow when at 0% science, or when at 100% science in the case of gold specialized cities.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
jar2574 said:
Even better then. It lets you wait for someone to discover the tech and trade for it instead. This is certainly better since you said it was more efficient. Then research the next tech at blazing speed since you can run 100% science for quite some turns by then.
jar2574 said:
The point is that you're not in the middle of researching a tech now. Since you were running 0% science, you're maybe at 5/2000 researched. There's really no penalty to switching!
Say that someone else has discovered the tech. By your own admission, it would now be more efficient to trade for it and research something else instead.
Can we explore it a bit further? (Of course, I agree with you that specialists are pathetic generally without Representation, so I should add that I mean to have representation as well.) My idea is something like the following:Zombie69 said:
let's suppose you have a city with a library that currently (at the stage of 100% science) is producing 4n + 3 beakers (just loses the fraction for the library effect), so gets n more beakers from the library. If we remove a worker from a "1 commerce tile" and make him a scientist (under Representation) it will produce 4n+8 beakers, so will also get n+2 beakers from Library. In other words, in this case the benefit of a scientist isn't the usual 6 beakers but 7 (and this is quite big number, for all standards). When the circumstances will change (for example, a hamlet has grown) then you reassign workers as fit.
Something similar can happen when in 0% science in a city with a market and 4n + 2 commerce (plus n gold from Market): again removing a worker from a 1 commerce tile and making him a Merchant results in a city with 4n+4 (plus n+1) gold, where with Representation the "net gain" is 7 "gold+beakers".
I believe that the two things to consider are food and cottage growth - but it definitely doesn't look to me very easy to say that this doesn't seem profitable.
EDIT: I just noticed that the Merchant effect is 6 (my visual math seem very bad), so it doesn't work (of course, because it creates a "lost fraction" 3 on beakers). Only the scientist gets 7.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
Like i said, "still, sometimes i will do this". Most of the time it's not profitable (because specialists don't produce much), but in rare circumstances it is.
Note that i would still prefer working a new cottage at 2 food + 1 commerce than a scientist, even if he gives 7 beakers. Assuming of course, that the GPP are useless in this city. For one thing, i usually value a growing cottage/hamlet/village as +2 commerce over what it currently produces, to represent the fact that assigning a working to it lets it grow. For example, a 2 food 3 commerce cottage (river and financial) to me is equivalent to a 2 food 5 commerce tile that won't grow. Also, don't underestimate the power of the 2 food, which is huge.
Note that i would still prefer working a new cottage at 2 food + 1 commerce than a scientist, even if he gives 7 beakers. Assuming of course, that the GPP are useless in this city. For one thing, i usually value a growing cottage/hamlet/village as +2 commerce over what it currently produces, to represent the fact that assigning a working to it lets it grow. For example, a 2 food 3 commerce cottage (river and financial) to me is equivalent to a 2 food 5 commerce tile that won't grow. Also, don't underestimate the power of the 2 food, which is huge.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
atreas said:
Note that with a library and a monastery, you get 35% bonus, which is very close to 33.3%. This means that you normally want multiples of 3. This makes representation more interesting.
The actual rule of thumb for a 35% bonus is that you want 3n, and -1 for every 21 (i.e. -1 at 21 making it 20, -2 at 42 making it 40, etc.) So the numbers to aim for become 3, 6, 9, 12, 15, 18, 20, 23, 26, 29, 32, 35, 38, 40, 43, etc. The only numbers without any fractional loss at all are multiples of 20, but the other numbers given above provide little fractional loss.
Zombie69 said:
No, binary science does not "let me wait for someone to discover the tech and trade for it instead." I already do that.
You claimed that "binary research would still be useful for the extra bonuses obtained from civs knowing the tech." Those bonuses won't come into play if I don't research the techs that civs already know.
People who use binary could choose to research techs that the AI is researching. Or they could do as I do and let the AI research the techs and then trade for them. People who do not use binary can do the same thing. Non-binary researcher may research techs more slowly, but if a binary researcher is wasting time researching techs that the AI has instead of trading for them then the binary researcher could still fall behind in the long run.
Zombie69 said:
Yes. I will try binary because it sounds like it may be faster than my old way of research. If it is faster then I will be able to trade a tech with the AI earlier than I would have without binary. But it will not speed up AI research. And it has nothing to do with whether or not I research techs that the AI is researching.
Zombie69 said:
Obviously there is no penalty to switching if you haven't started researching something, regardless of whether you use binary or non-binary research. My point is that the opportunity to switch tech research while having commerce at 100% is irrelevant and is not an advantage to using binary if you never switched tech research when you are not using binary.
So that 'advantage' seems to be overrated, unless the player is someone who changes their research goals.
Zombie69 said:
My point is that this is also an overrated 'advantage' of binary, because except for meditation and priesthood, it is extremely rare for anyone to have discovered a tech that I am in the midst of researching. I typically research BW and then beeline to ABC. At that point I go for code of laws and civil service. If I can trade for meditation and priesthood then I do, and if no one will trade them then I must research them.
After that, I research "advanced" techs and let my trades with the AI back fill the techs I've missed. You can see why it would be rare for anyone to be research the tech I'm currently researching.
My point is not to say that binary sucks, it's to say that some of these 'advantages' are really overrated.
Zombie69 said:
Zombie69 said:
So in a city with cottages, you look for multiples of 4, but in your head you are also considering the value of growing the cottages? This would be difficult and time consuming. I'm not saying that this couldn't be done, but I think it's putting too much emphasis on the multiple of 4 part and not enough emphasis on the potential value of the cottages.
Instead of worrying about the multiples I'll just keep developing each tile in order of its value. (i.e. floodplains first, river grassland second, other grassland third.)
In non-cottage cities I think you raise a valid point.
Zombie69
Emperor
- Joined
- Nov 22, 2005
- Messages
- 1,902
jar2574 said:
You don't seem to understand. Let's say that you'd like to research a tech at 80% and it would take you 10 turns. By instead doing 0% for 2 turns, then 100% for 8 turns, you gain the extra flexibility of changing your tech to research after 2 turns. If you had gone for 80% all the way, you'd already have 20% of the tech researched after 2 turns so switching would not be as feasible. I'm not saying you'll use this flexibility everytime, but having it is nice. After all, there's no downside.
jar2574 said:
You can't do this all game, that's for sure. There's an undocumented limit coded into the game of how many techs you can obtain through trade. Once you reach that limit, the AIs won't trade with you anymore because "we feel that you are becoming too advanced", even if they're 10 times as advanced as you are (so the text doesn't reflect the actual reason, but i guess the developpers wanted to hide the actual reason and that's what they came up with to hide it).
Therefore you can't rely on trade alone, and at some point you'll have to research techs competitively with the AI.
