Optimum Early Game Research

slaze

Deity
Joined
Nov 17, 2006
Messages
2,287
Location
california
Optimum early teching works like this: you want to spend as much time as possible at 0% science.

Many of you probably laughed that off as a joke, but hear me out... Now, most people tech at the highest rate they can. If, after maintenance costs, the highest they can go is 80%:science:, then 80%:science: it is. But its perfectly equivilant to run 60% half the time and 100% the other half. So if this is the case, you'll want to be running some combination of 100%:science: or 0%:science:, and here's why.

Now it's true the following corner-cutting moves don't save that much in terms of research accumulated. It might save you something like 1 or 2 turns or even less for the whole of the ancient age. But it does save some, and that makes it slightly better. And better is in the direction of optimum. Anyway, read on.

The only constant, as far as your economy is concerned, is maintenance. You can choose to reseach, you can choose to put commerce into espionage, you can choose to stockpile gold. But outside of not founding cities (or running certain civics, etc., etc.) you can't choose whether you have to pay maintenance - it's a must.

So here's where the tricks come in. Say you want Monarchy (a good economic tech that is worth breaking your bank on). But right now you don't even have Meditation. You're not concerned about Poly or Mono, you're going to take the shortest path: Meditation, Priesthood, Monarchy. What you do is you put the science slider at 100%:science: and you go into the scientific advisor screen and count up how many turns it would take to get all three techs (say you find out it's 20 turns). You then put the slider down to 0% to see how much maintenance you have to pay each turn (say it's 15:gold: per turn).

What you'd do is run 0%:science: until you get 300:gold: (or a little bit less as you'll see) and then run 100%:science: all the way to Monarchy. There's a couple reasons why this is better.

First, your economy is growing. Your cities are adding more population allowing more commerce tiles and your cottages are growing into hamlets and such. If you have more :science: multipliers which, with the early library is usually the case, you'll have more commerce to put in the multiplier and while you're running the best rate, 100%:science:.

The other reason is that while running 0%:science:, you still get 1:science:. While running 0%:gold:, you don't still get 1:gold:. In the above Monarchy example, this might save you 16. Yes that's not much, but in the very beginning this is about as much as you'd make in a turn. So you save only a turn, but it's closer to optimal.

Another thing to do is to use the hidden multipliers. If you're researching Writing and have all the prereqs (+20% each for +60%) and can afford to delay Writing for a turn, use the prereqs to boost research into the next tech. Say you want Alphabet which only has Writing as a prereq, then research as much of Writing as you can without completing it, and then on the next turn go 100%:science: and pour as much of the multiplied research into the next tech. That research gets a +60% modifier, while then next turn Alphabet would go back down to a +20% modifier. After that you would go back to the 0%:science:, save, 100%:science: model. Use techs with many prereqs to spill research into techs without.

To conclude, for a growing economy, the 0%->100% model is best. For a shrinking economy (losing cities to invaders) a 100%->0% is best. For a stagnant economy, it doesn't matter. The 0% one beaker trick dies as soon as you run your first scientist, although as the techs (and your economy) get bigger, the 1:science: doesn't really matter, although you could switch to merchants if you really wanted to milk it.

And I could see a weakness in this as a vulnerability to spies, with their steal treasury mission.
 
And like this you're missing possible trades that AI's could offer. Thus losing way more breakers then you originally thought you're winning. :mischief: Also how does this work in combination with rushing, whipping, warfare.. which influences your commerce. Good luck trying to calculate everything.

Also AI will demand your gold and on high difficulty you don't want to take diplo penalties like that. There is a big difference with the AI asking a tech you already have or gold that you saved up to get a tech. :rolleyes: If you give the tech, you still have the tech yourself. If you give the gold .. you're fcked. :lol:

Imho in theory it works and makes sense, in practice it doesn't.
 
The 1 research per turn is certainly a reason for the 0/100 pattern, albeit a bit too micromanagey for me

The growth argument is only partially true, in the sense that you are still buildiing libraries. If you are building a bunch of libraries/unis/obs, save up the money and research after they are done. The reverse holds for markets/grocers/banks, if you generally have a money reserver so you can start by 'borrowing' from that reserver.

Support you need 750 beakers, and already have libraries everywhere, and 10 gpt maintenance. Your commerce is 20 in the first 20 turns and 50 after that (ie economic growth) while maintenance is constant (it would go slightly up in reality).

Doing maximal research:

20 turns at 50% science (since 10gpt maintenance = .5 * 20 commerce) = 250 beakers (20 * 10 * 1.25)
10 turns at 80% science (since 10=.2*50) = 500 beakers (10 * 40 * 1.25)

Net result: 30 turns, 750 beakers, no net gold change

Doing 0/100 research

0% science for 15 turns = 150 gold in treasury
100% science for 5 turns = 5 * 20 * 1.25= 125 beakers, - 50 gold
100% for 10 more turns = 10 * 50 * 1.25 = 625 beakers, - 100 gold

Net result after 15 turns in a growing economy: 750 beakers, 0 net gold


What I wanted to say is: a growing economy (commerce per turn increasing) does not give innate benefits to a 0/100 pattern.

There are valid reasons for a 0/100 pattern, these are:
- 1 beaker per turn. Although in the time frame this really matters you will likely have low maintenance, so it doesn't matter that much.
- information (if you only want theology for the religion, it is good to first stock cash, and only start researching if you can finish at 100% science, since you want to stop if an AI discovers it first and this way you have minimal investment in the tech). If an AI suddenly DOW's you, you might want to stop researching literature and focus on the war. If you do 100%/0%, there is a good chance that you have wasted less beakers on literature before the DOW and have some cash to burn on upgrades / maintenance / bribes / teching military
- science multiplier from known civs already knowing the tech
- upcoming science multipliers from library, university, academy. Ie as soon as you discover writing or education, set slider to 0%, build the buildings, and set slider to 100% to burn the cash at profit. If you have a great scientist coming up in a reasonable time frame that you plan on using for an academy, do the same.

Due to the extra micromanagement, I only do this after writing, education, and astronomy, waiting for my big science cities to build their infra.
 
Interesting. I've used a more limited version of this before - if I didn't need immediate techs (resource-related mostly), I would stockpile gold then switch to deficit research once I build libraries in my larger cities.
 
Most of the tricks and ideas Zombie69 raised in that micromanagement article were fixed in patches to vanilla and Warlords and don't apply at all to BtS.

There is some value in the idea of saving gold and then using deficit spending to use that gold for fast research. It mostly depends on the distribution of science multipliers, if they will increase a lot in the future then save gold now and research later with the improved multipliers. That can be worth several hundred bonus beakers.

Several early game situations spring to mind where this trick can really pay off. Say you have Writing and four cities. The capital has built the library and is running 2 scientists for an academy. The three other cities are all building their library. Saving gold while running the scientists in the capital will build up a nice stockpile. Then when the academy and 3 extra libraries are finished raise the research slider and gain the benefit. A similar opportunity occurs when you build universities in many cities and later observatories.

Another benefit of delaying research spending is the fact that you gain a small discount on researching a technology if another civ that you have met has acquired that technology. If you wait for the other civ to finish researching a tech you will pay slightly less to research it yourself. The amounts saved are just a few % but it all adds up and some techs can be delayed, so wait for the discount.
 
Most of the tricks and ideas Zombie69 raised in that micromanagement article were fixed in patches to vanilla and Warlords and don't apply at all to BtS.
Most , but not all... it is still worthwhile to beeline a tech line and then resume the other lines because of the discout to to research of techs that other civs have... it is worthwhile ( as slaze said ) to research in bursts because of the roundings and because of the "surprise" effect ( AI do not cope well human bursts in research either via deficit spending or via bulbing )...

Zombie article has some good ideas, but as you said , it is completely outdated in most issues...... but it is still worth a reading.
 
Doing maximal research:

20 turns at 50% science (since 10gpt maintenance = .5 * 20 commerce) = 250 beakers (20 * 10 * 1.25)
10 turns at 80% science (since 10=.2*50) = 500 beakers (10 * 40 * 1.25)

Net result: 30 turns, 750 beakers, no net gold change

Doing 0/100 research

0% science for 15 turns = 150 gold in treasury
100% science for 5 turns = 5 * 20 * 1.25= 125 beakers, - 50 gold
100% for 10 more turns = 10 * 50 * 1.25 = 625 beakers, - 100 gold

Net result after 15 turns in a growing economy: 750 beakers, 0 net gold


What I wanted to say is: a growing economy (commerce per turn increasing) does not give innate benefits to a 0/100 pattern.

Yeah I guess I needed to hear this. The deficit phase needs the surplus phase, so any commerce gained during the deficit phase is balanced out by not being there in the surplus stage.

So I guess the only time this works, from a multiplier point of view, is stockpiling gold and then running deficit after your multipliers are built.
 
Also AI will demand your gold and on high difficulty you don't want to take diplo penalties like that. There is a big difference with the AI asking a tech you already have or gold that you saved up to get a tech. :rolleyes: If you give the tech, you still have the tech yourself. If you give the gold .. you're fcked. :lol:

Imho in theory it works and makes sense, in practice it doesn't.

Except that he specifically says this is for early-game.

Until you or the demanding AI researches Currency, they cannot demand money from you anyway, so that's a non-issue for the early-early game.
 
Except that he specifically says this is for early-game.

Until you or the demanding AI researches Currency, they cannot demand money from you anyway, so that's a non-issue for the early-early game.

Early game for me is pre-lib, he's talking about monarchy here. That's on the same time line as currency and after alphabet to say at least. Hence you're missing partial tech trades that AI's could offer and you risk losing your gold. (or getting diplo penalty)

Also he talks about spies, so I think I can assume that he ALSO thinks early game = pre-lib. :)
 
Early game for me is pre-lib, he's talking about monarchy here. That's on the same time line as currency and after alphabet to say at least. Hence you're missing partial tech trades that AI's could offer and you risk losing your gold. (or getting diplo penalty)

Also he talks about spies, so I think I can assume that he ALSO thinks early game = pre-lib. :)

I don't understand how you would miss tech trades, surely they could still make them while you run deficit. The accumulation of gold may even affect trades by having them offer trades where your gold takes place, sorta the opposite of partial tech trades.

As far as worrying about diplo penalties, well, is it really best to be broke all the time? Surely one can feel out the diplomatic situation. A pangea's different than a continent with Hatty and Gandhi. I play multiplayer mostly anyway, so this is of much more importance to you.

And by early game i mean when the single :science: gained by running 0% has the most impact. That may be pre-Writing, that may be pre-CS, that may be pre-lib. Although the upper side of pre-lib might be too much.
 
I don't understand how you would miss tech trades, surely they could still make them while you run deficit. The accumulation of gold may even affect trades by having them offer trades where your gold takes place, sorta the opposite of partial tech trades.

As far as worrying about diplo penalties, well, is it really best to be broke all the time? Surely one can feel out the diplomatic situation. A pangea's different than a continent with Hatty and Gandhi. I play multiplayer mostly anyway, so this is of much more importance to you.

And by early game i mean when the single :science: gained by running 0% has the most impact. That may be pre-Writing, that may be pre-CS, that may be pre-lib. Although the upper side of pre-lib might be too much.

It's quite simple, let's say the AI wants for example maths from you. He has alphabet but maths is lower valued. You're researching alphabet, there's a chance he might offer the trade. Not only does this not count to wfyabta but you get the tech you wanted as well. And this happens quite frequently in my games. With other words if you were just piling up gold, you would never had that deal.

You can of course argue that the same deal would have been offered once you actually start teching .. after piling up gold. But let's use the same example. If I were teching partially alph and the AI started to tech maths, I could possibly still get the deal. In the 100% gold approach, by the time you start teching the AI will have more turns in maths already and the deal is no longer valid.Maybe this sounds farfetched but the whole aim of this thread is to get 1-2 turns advantage but if you would miss such a trade, there's no way this method can make up for those lost breakers. That's how I see it at least.

And let's be honest, no one "buys" techs of the AI. Afaik people try to avoid putting gold in trades because the AI values it so low.

I noticed AI's demand gold a lot if you have a lot of it. Almost every time I do trade mission with GM in the next 10-20 turns some idiot will demand a ridiculous amount of gold from me. Maybe I'm just unlucky but I think it's a pretty big argument against this approach.

edit: fixed typo
 
It's quite simple, let's say the AI wants for example maths from you. He has alphabet but maths is lower valued. You're researching alphabet, there's a chance he might offer the trade. Not only does this not count to wfyabta but you get the tech you wanted as well. And this happens quite frequently in my games. With other words if you were just piling up gold, you would never had that deal.

Run that one by me again?
 
It's worth noting that this is most effective on Quick, and least effective on Marathon.
 
Run that one by me again?

[offtopic]

I know you have a thread on that and I did read it. Actually I thought I got that idea from your thread but apparently not, I wonder from who I have that.

But I also know your thread isn't 100% correct. E.g. wfyabta does not affect friendly AI's, which some of my previous games have proven wrong (no vassals or whatever). Very high diplo yet wfyabta.

I'll let this issue just stay in the middle but how would you explain trades like this? :





Anyways even if it were wrong, it doesn't make my point on this topic less valid.
 
A very important point the OP has missed is that there is a significant difference between hardcoded :science: cost and the effect of :science:

Because, basically, three factors should generally be increasing throughout the game;
1 Maintenance (negative)
2 Commerce output
3 expected/required commerce output

So saving gold enough to run that exact amount of turns at 0% might prove to be a bad move as the maintenance will eat the gold faster the longer you wait. Also, consider Inflation as well as there are plenty of random events that require :gold:, so regularly hitting 0 is bad.
 
as the maintenance will eat the gold faster the longer you wait.

I don't think I understand you correctly. Maintenance has nothing to do with gold reserve, nor does inflation. Yes maintenace increases with pop but usually with over-compensating increases in commerce, as a 0.2:gold: cost in pop maintenance is covered by a 1:commerce: from a new river tile.

Other than Burn's points, which are immortal issues and have nothing to do with the commerce engine, it does no harm to run this sort of research.
 
Most techs give an immediate boost to your research rate, directly or indirectly.

So the optimal play is zeroing your treasury every time you discover a useful tech.
 
Back
Top Bottom