Binary Research

[CarlH]

Hi all,

Long time lurker, first time poster.

Like a lot of people's first comments on the forums, I'd like to thank everyone, particularly the regulars, for all the time they put into this forum and how much it has helped my play. I have moved from Prince to Immortal in the last 6 months, mainly due to the wealth of Knowledge available here. I have only won one game on Immortal and that was as HC of the Incans, so it doesn't really count. But I'll keep on trying.

Anyway, I adopted the binary research idea since I started on Emperor and can appreciate that it helps because of the CIV's rounding down mechanism. However, setting the science slider to either 0% or 100% means that either your commerce multipliers or science multipliers have no effect part of the time (putting aside specialists and shrines/corps headquaters). I have recently been wondering whether this factor makes binary research less effective?

This is a maths question and I am probably being lazy asking, but I know there are some in the forums who love the calculation side of CIV, maybe they can help.

Thanks in advance.

 

[Sero Sed Serio]

I just recently posted a thread under the exact same title, however my question was slightly different.

You use binary research, so you already understand that it is a powerful tool and that any inefficiency is greatly made up for. It seems like what you're really asking is "Can someone break down why this works".

I'm curious myself. I do understand that building up a lot of cash and then spending it immediately after you get libraries in a couple of cities can pay off. I have been wondering how that excess cash figures into things though. Does it filter through your cities or is it directly added in a 1:1 to your research? Because if it was filtered through your cities, with their modifiers, it would seem obvious that modifying your commerce once (through gold modifiers) and then again (through beaker modifiers) would be very powerful.

 

[CarlH]

To be honest, I had taken the advice of superior players that it worked, rather than being sure that it did.

It has recently struck me that, particularly in the case of later expensive techs, having research at 0% for say 10 turns seems to be a lot of time during which libraries, universities, observatories and oxford are having no effect (and obviously vice versa for markets, grocers, banks and wall street).

Also, I thought that it was to be applied not only when building the modifiers, but for every tech: ie artillery will take 10 turns at -200 gold, so i should turn slider to 0%, save up 2000 gold, then set slider to 100%. rinse and repeat.

 

[JujuLautre]

All commerce multipliers have the same effect.

Let's simplify and take a tech for which you need 10 turns. You break even at 40% science (60% gold). Let's take a city having 20 commerce and a library.
If you run at 40% science for 10 turns, the library will net you an extra 20*0.4*0.25 = 2 beakers per turn; so 20 beakers during 10 turns.
If you run at 100% gold for 6 turns, and then 100% science for 4 turns, the same library will net you 0 beaker for 6 turns, and 20*1.00*0.25 = 5 beakers per turn for 4 turn, so 20 beakers during these 10 turns.

In the end, the result is the same. But obviously, this is for commerce multipliers being here from the beginning to the end of the process. As stated before, if they appear midway it could be better to emphasize science or gold for a while just for that.

 

[VoiceOfUnreason]

Also, I thought that it was to be applied not only when building the modifiers, but for every tech: ie artillery will take 10 turns at -200 gold, so i should turn slider to 0%, save up 2000 gold, then set slider to 100%. rinse and repeat.

Yes

It has recently struck me that, particularly in the case of later expensive techs, having research at 0% for say 10 turns seems to be a lot of time during which libraries, universities, observatories and oxford are having no effect (and obviously vice versa for markets, grocers, banks and wall street).

To first order, this is true - the research multipliers are idle in cities with no research, the weath modifiers are idle in cities with no income.

But while true, it's not particularly relevant. 400 base research with a 25% modifier is 500 beakers... which can be 500 beakers per turn for one turn, or 50 beakers per turn for 10 turns. The net result is the same.

The major difference between the two is that once you've chosen to invest 50 beakers are turn 1, you can't change your mind on turn 2 to get the gold back. However, converting the commerce to gold means that, if you change your mind, you can recover the research by bumping the slider up, and burning some of your surplus treasury.

Thinking in broader terms - consider the situation where you are considering which of two 500 beaker techs to research. Investing in research immediately means that changing your mind resets the clock (in the previous example, each tech is 10 turns away; if we put 3 turns into the first tech, then change our minds, the second tech is still 10 turns away. If instead we build up the treasury - we are effectively getting 9 more turns of information before we have to choose which tech to research, and that extra information is "free".

One of the things that can happen during the binary window is that you get an opportunity to trade for the "other" optional prerequisite technology - meaning another research discount.

Beyond that, there's the additional possibility that the information we get tells us we'd rather spend the treasury in some other way (upgrading units, buying some tech, bribing the AI into war, etc). So we also get added flexibility, again for "free".


There are some additional imbalances that you can play with here. The first, as noted, is that if you've just discovered a tech that unlocks a bunch of research multipliers, turning off research while you construct buildings effectively banks the effects of the multiplier during the turns you are building.

Likewise, you want to be in research mode when constructing the banking multipliers, but notice this requires a bit more thought (if we burn the treasury to zero to discover banking, "turning off gold while we build" is a problem).


In addition, there's a bit of zim zam that can be played with the cities themselves, and the way they develop. If you want to accelerate growth, or production, in a city, you need to sacrifice commerce to do so. The opportunity cost of this is lower when the slider is pointing the "wrong way". So if you just got a new happy resource, you can score a small profit by turning the slider to gold, growing the research cities, cranking the slider to research, grow the gold cities, return to normal rhythm.


There's an additional quirk in the game, that techs become cheaper to research when more civs have discovered them. So there's a discount available on the research cost incurred after a neighbor discovers the tech you are researching (I'm not actually sure of the precise mechanics here, beyond the point that - like pre-requisite bonuses - the multiplier is applied after all the city research is pooled and the bonus beaker is added).

Summed up in one single lump, research bonuses are never (?) lower at the end of a research cycle than they were at the beginning, but they are sometimes higher. Queuing all of your beakers for the end of the cycle ensures that you push as much research as possible through the bigger multiplier without sacrificing time.

 

[T-hawk]

Very well said, Voice. Binary research is about both flexibility (delay the irreversible decision to invest beakers) and mathematical optimization (produce beakers when their multiplier is highest.) Another reason to keep cash for flexibility is random events that may demand a sum of gold.

To address a point from the original poster: Yes, binary research means that sometimes your multiplier buildings will go idle. But that also means that the opposite multiplier is working on your entire commerce production. When science is off, Wall Street in your 100 commerce city is working on all 100 commerce. Wall Street working on 100 commerce half the time is the same as Wall Street working on 50 commerce all the time. But if you can optimize those times so that the science time occurs during periods of higher beaker multiplier (later), you come out ahead.

Summed up in one single lump, research bonuses are never (?) lower at the end of a research cycle than they were at the beginning

I can think of a few ways to lower your research multiplier bonus. One is to change civics from Free Religion to something else. Two is when monasteries expire, although that would only happen during a research cycle if you traded for Scientific Method. Another is to lose a city with a high research multiplier: suppose you had three cities with library, library, and library+academy+Oxford. Your average research multiplier is +75%. Then if the Oxford city gets razed, your average research multiplier drops to only +25%.

All those are pretty uncommon though.

 

[Grey Fox]

The best ways to use Binary research is to save up gold while your research multipliers are low and then increase it to max once you got the research multipliers in place. For example the turn you get Writing, build libraries and turn the slider up again once you either have libraries or an Acadamy.

Once you're at Education, same deal. Or while building Oxford.

Other reason is to get a tech quick in one research cycle to trade it. This is to be sure to have something to trade. If you are teching it at a low slider someone might get it before you do and then much of your research was a waste (unless you have something to trade for the rest of the tech).

 

[RedFury]

To give a solid example which actually happened to me last Emperor game:

I had done a semi-rex and was recovering my economy and was breaking even at about 20% science slider. At this tech rate, I could research either aesthetics or alphabet in about 30 turns. I thought I would probably go Alphabet because I was confident at beating all the AI's to it.

I used binary research, meaning at 20% break-even rate I would run 0% for 24 turns then 100% for 6 turns. Well ... after 22 turns one of the AI's got alphabet and started trading it around. With binary research I was still 2 turns away from switching to research so I was able to change my tech choice from Alpha to Aesthetics with no penalty. By turn 30 when I finished Aesthetics pretty much all the AI's had alpha, so had I not used binary research, I would have been stuck with a 3/4 researched tech which wasn't any good as trade-bait to catch up in tech.

I hope this makes sense

 

[BarrageQueen]

To be honest, I had taken the advice of superior players that it worked, rather than being sure that it did.

It has recently struck me that, particularly in the case of later expensive techs, having research at 0% for say 10 turns seems to be a lot of time during which libraries, universities, observatories and oxford are having no effect (and obviously vice versa for markets, grocers, banks and wall street).

Also, I thought that it was to be applied not only when building the modifiers, but for every tech: ie artillery will take 10 turns at -200 gold, so i should turn slider to 0%, save up 2000 gold, then set slider to 100%. rinse and repeat.

Your point is that turning research on/off binary style renders one of the research/gold multipliers null. Well, assuming that your economy had the same multipliers for both research and gold, commerce flowing through binary research or not wouldn't make a difference in the total outflow. In practice, the multipliers are different and they get built at different times, so that's why people are mentioning the build wealth before a library/uni/observatory gets built, etc. So there's really no "wasting" of commerce multipliers, from the binary division.

 

[Phatkarp]

So what I'm hearing is that the benefit of binary research is twofold:

1) You occassionally get more benefit from the research multiplier buildings completed during the cash phase;

2) You get the benefits of delaying your research decision (which could be information re: what other AI's have researched, or trading for backfill techs which make the target tech cheaper).

 

[Kesshi]

2) You get the benefits of delaying your research decision (which could be information re: what other AI's have researched, or trading for backfill techs which make the target tech cheaper).

The only fault of number 2, is that it becomes more useful the lower your tech rate would be.

That is, if you can learn your next tech in 20 turns at 80% science and 20% gold, 4 turns of gold isn't going to show you much. However, if you can learn your next tech in 20 turns at 20% science and 80% gold, 16 turns of gold could be very informative.

This isn't always a bad thing, as I am one of the most vocal people to bark at someone who says that they don't like to run under 50% science, or someone who asks "what is a good tech percentage," etc. But to make the most out of number 2, it does require a under-normal-circumstances-lower-science-slider situation.

 

[madscientist]

Binary research is very powerful and useful. I use it often at emperor level and have no regrets. There are also games where research is adequate and I prefer NOT to use it.

 

[nastyjames]

What about when you need a few beakers of a tech before the AI will trade with you?

Also do you guys run this at all times or just on certain techs?