How science works and why there IS a 4 turn cap

[regoarrarr]

Okay, so I spent some time this evening trying to figure out how exactly the techs per turn works, and to put to rest the 4 turn per tech myth.

I think I have the formula figured out - here goes:

First, you need to calculate the base number of "beakers" needed to research a tech.

base# beakers = 16 * tech cost

where tech cost is the cost listed in the editor. I will post a follow up message in this thread listing off the various costs for those of you too lazy to look them up

Then there is an adjustment based on the number of civs that already have that tech. Multiply the base number by

(# of civs that do not have the tech) / (total # of civs in the game)

That's the number of beakers you have to get to get the tech.

Now let's talk caps. As most everyone has already figured out by now, there is a max cap of 32 turns / tech.

But, let's talk about the minimum cap of 4 turns / tech. Some are saying there is a cap, while others (myself included) have said that we have researched techs in 3 turns. But yet, I think both sides are right!

There is a 4 turn / tech cap only when dealing with the adjustment due to the number of civs that have a tech. This is a bit confusing, so let's do an example.

Mysticism has a tech cost of 4, so it costs 64 beakers. I was playing on a tiny map with 4 civs total, and the other 3 civs already had Mysticism. So, we multiply by 1/4 (# of civs that don't have the advance / total # of civs). That gives 16 beakers. I was producing 7 beakers, so it should have only taken 3 turns, yet it was capped at 4 turns. Had I been producing 22 beakers, that would give me 64/22 = 3 turns / tech.

I didn't test that, but I KNOW that I have gotten 3 turns / tech, so I'm fairly confident that that is what is going on. So, there is a cap when your number of beakers needed is adjusted due to other civs having the tech, but not when you can get it in under 4 even without this adjustment.

A couple of other things - the beakers you produce only go to the advance that you're currently researching. For instance, if you're researching Writing and you've produced 40 beakers towards it, but then you switch to Pottery, those 40 beakers are wasted. They don't transfer. You can switch while you're on the Science Advisor page (F6), but once you leave that page those beakers are gone.

One last thing - let me use an example. I was researching Polytheism, tech cost of 12. So the number of beakers needed is 16 * 12 = 192. With 7 beakers it would take 28 turns, 1-6 beakers we hit the cap of 32. I researched 6 turns at 7 beakers. That leaves 150 beakers to be researched. At this point, 7 beakers it takes 22 turns, 6 beakers it takes 25 turns, and with 1-5 turns it takes 26 turns. That is because we've already spent 6 turns researching this advance, which are taken out of the 32 turn max.

Anyway, this is what I have so far.

[regoarrarr]

Here's a list of tech costs in the order they are in the Civ Editor, which is more or less in the order they go on the tech tree, from Ancient to Modern:

Bronze Working	3

Masonry	4

Alphabet	5

Pottery	2

The Wheel	4

Warrior Code	3

Ceremonial Burial	2

Iron Working	6

Writing	8

Mysticism	4

Mathematics	8

Philosophy	6

Code of Laws	10

Literature	10

Map Making	12

Horseback Riding	5

Polytheism	12

Currency	16

The Republic	28

Monarchy	24

Construction	20

Monotheism	30

Feudalism	26

Engineering	36

Theology	32

Chivalry	30

Invention	40

Printing Press	32

Music Theory	30

Education	42

Gunpowder	44

Banking	48

Astronomy	48

Chemistry	52

Democracy	56

Economics	44

Navigation	52

Physics	56

Metallurgy	52

Free Artistry	56

Theory of Gravity	60

Magnetism	60

Military Tradition	56

Nationalism	100

Steam Power	96

Medicine	84

Communism	100

Industrialization	88

Electricity	100

Scientific Method	96

Sanitation	80

Espionage	76

The Corporation	92

Refining	100

Steel	100

Atomic Theory	160

Combustion	120

Replaceable Parts	84

Flight	140

Amphibious War	92

Mass Production	100

Electronics	120

Motorized Transportation	100

Advanced Flight	140

Radio	140

Rocketry	160

Fission	220

Computers	180

Recycling	240

Space Flight	240

Nuclear Power	200

Superconductor	220

Miniaturization	180

Ecology	160

Synthetic Fibers	200

Satellites	200

The Laser	220

Genetics	240

Stealth	220

Smart Weapons	160

Robotics	200

Integrated Defense	180

[SandMonkey]

is it really this simple? it doesn't seem possible, considering it took so long to figure it out for civ2 (no one ever did completely figure it out, did they? I think they figured out too many variables).

[David Weldon]

I found that the 16 you use is map-size dependant. On standard it is 24, and on small it is 20. I believe the number comes from the number in the editor for world size, divided by 5. (i.e. normal: 120/5 = 24, standard: 100/5 = 20). I assume your tests were done on a tiny map? If so, then this would agree with the two data points I have already taken.

As for building up tech, you can get the tech either by spending 32 turns, or by getting enough beakers, whichever comes first. That's why the first few turns, even though you're being limited by the 32 turn rule, you are still accumulating beakers that help if you science rate goes up in the future.

As for the 4 turn soft limit, I believe the penalty is worse than you describe because I have been able to get a new tech (no one else has it yet) at 40% science in 4 turns. at 100% science, it is still 4 turns. This means that more than double the # of beakers doesn't even reduce the time from 4 to 3, much less 2. There is definitely a steep penalty here, but I haven't been able to find exactly what it is because I have never been able to get a 3-turn advance (I'm not saying it's not possible, only that I haven't been able to test it yet).

[Kc7mxo]

(# of civs that do not have the tech) / (total # of civs in the game)

so, what you're saying is that the fewer civs that have a tech, the easier it is to discover? umm. are you sure?

[Grunthex]

No, he's saying the MORE civs have it the easier. For example if 6 out of 8 civs on a standard map have it, the cost would be:

(TechCost*24)*(2/8)

The 2 being the civs that DON'T have the tech, the 8 being the total, and the 24 being the modifier for the map size.

Question: If 2 of those 8 civs get eliminated, does the denominator for techs after that get reduced to 6, or stay at 8?

[regoarrarr]

That is correct, the more Civs that have it, the easier it is.

Also, to David Weldon, yes, I used a Tiny map, so it looks like you are correct in figuring that it is map-dependent.

Another thing that I forgot to add last night is that if a civ discovers the advance you're researching while you're in the middle of it, it automatically reduces the number of beakers you need.

For example, I was researching Iron Working and I had 5 turns left to discover it. Next turn I automatically discovered it - turns out that one of the other civs had discovered it, so I got it automatically because the number of beakers I had already researched was greater than the new number I needed to get, taking into account the 3/4 multiplier now that one other civ had it.

[Dragon]

This means that it is very important, if you sell or trade a tech that you have just discovered, immediately to sell, trade, or give it to everyone. Otherwise, even if the AI doesn't trade for it, it becomes very cheap to research.

It's tempting, if the last AI civ offers to pay 20 gold for a tech that you just sold off to everyone else, to say "Forget it." Perhaps a better strategy is to give them the tech. They might be unexpectedly grateful.

[Andy]

It is also important in trading not to accept a tech that you are currently working on if there is another option available. During my last game I made a trade to recieve my current research choice with only a few turns left on it when there was another tech I could have gone for. Basically I wasted all of the beakers that I had used to get the tech. This also means that switching research can be very costly.

[Achnor]

I think that if you switch research-focus the beakers you have used on that tech _stays_ with that tech, so if you vchange back later you can conytinue where you left...

Achnor

[nickersonm]

there is a "tech cost" or "tech rate" modifier in the map-types. This must be the multiplier.

- nickersonm

[regoarrarr]

No - if you switch techs that research is gone forever. Research a tech for a few turns, then switch and leave the screen. If you go back to the screen and switch back to the first tech, it will still be at 32 turns (or however many turns it would normally take)

[La Fayette]

Originally posted by SandMonkey
it doesn't seem possible, considering it took so long to figure it out for civ2 (no one ever did completely figure it out, did they? I think they figured out too many variables).

Yes, it took a long time, but it is completely clear now, after the last discoveries by samson (in april 2001 IIRC). You can find it in the GL (top thread of civ2 strategy forum). Look at 'beakers'.

[David Weldon]

I'm still looking for an exact measurement of how big the 4-turn barrier is. I _still_ haven't been able to get a new tech in 3 turns, even though it only takes 30% science to get it in 4 turns. Could anyone who is able to do this please post the # of beakers and # of turns for each % setting (along with world size and difficulty level)? Or should I start to disbelieve those who say they've done it?

It would be really good to have this info.

[KrazyHorse]

Don't disbelieve. I've gotten a few 3 turn advances.

[HalfLotus]

Nice work rego. Looks like a winner to me, and it matches my experiences. My problem is that this handicaps the player leading the tech race.
This makes it very difficult to get far ahead in the tech race. After I discover a tech, it is automatically easier for all other civs? I don't want the game to be easier, but I want my scientific efforts to keep me ahead in science, not be undermined by this formula. I think it blows.

[solo]

How about beaker carryover towards the next advance? Does it work the way it does in CivII, or are the extra beakers from the remaining cities wasted on the turn an advance is secured?

[Matthew]

So as I understand it, when you get a tech you have been researching, rather by trade or hut, all the beakers you have been saving up are wasted? Or do they perhaps go into the tech the computer automatically selects for you once you get the one you had been researching? It may be that they go towards the tech the computer selects for you.

[Matthew]

It seems that a tech lead isn't necessarily that hard to maintain, provided you are willing not to share techs with other civs. This would at least reduce the rate at which they can catch up to you. This will cost a lot in potential resources, but may be necessary if you want to assure a win.

Of corse it is also helpful to take quick advantage of a tech lead you have before it goes away. This is made more difficult by the reduced strength of more advanced units compared to older ones.

Slightly off topic, I think that more advanced units need to be stronger in multiplayer, in order to keep smaller, more advanced civs on a reasonable par with larger, less advanced civs.

[jonasf]

They are supposed to get cheaper techs..... I'm just wondering exactly HOW MUCH CHEAPER????

Also the cheaper city improvements.... How much cheaper are they?

Same with the religious civs... how much cheaper are the religious improvments?

[pchang]

Scientific civs do not have cheaper research costs. However, a lot of their advantages makes it easier for them to generate more beakers. Libraries and Universities cost 1/2 the shields (research labs too).

[David Weldon]

Wasted Beakers:
You're right Mathew, all excess beakers are lost. You can micro-manage and save some of these by reducing your science rate when you get to 1 turn left (this will convert a lot of beakers into gold that turn, but still allow you to get the tech in 1 turn). I wish they would automatically convert excess produced beakers into gold so that the player doesn't have to micro-manage like this, but they don't.

As for accumulated beakers (e.g. if you've been working on a tech for 10 turns, and then you trade for it or get a hut with it), they are irrevocably lost. I'm quite certain this will never be changed, as that's just the way it is. It's kind of a bummer, but really not that bad once you get used to it and plan for it. The up side is that if you've researched almost all of a tech, it costs much less in trade to "finish it up" if you want to.

Civ Strengths:
Scientific gives you cheaper science-related city improvements (like libraries, universities, etc...), not cheaper science advancement. It makes the shield cost 1/2 of normal for those city improvements. You also get a free advance when you enter a new age, but it is chosen by the computer and tends to be the cheapest one you could have gotten (in other words, don't skip techs or you'll get screwed). In general the increased growth of Industrial or the reduced corruption of Commercial each tend to yield faster science advancement than Scientific does. They kinda screwed that up...

Religious civs get 1/2 cost for religious improvements (temple, cathedral).

[inca911]

Wonderful work. Just goes to show you what a geek with a calculator and way too much free time can accomplish! So this equalization principle (the formula you describe is appears to be used to slow down the tech leader more than Civ2) that you have characterized now provides two new schools of thought:

1. The Science King: The goal of the science king is to maintain the technological lead in the game and trade new technology discoveries to all civs to feed your treasury and fuel your empire into the #1 civ on the planet. If you don't trade the tech, rest assured that your neighbors will!

2. The Science Buyer: The goal of the science buyer is to ignore scientific development costs and to simply purchase these on the open market. It certainly would be an interesting option to ignore science and focus instead on culture and economic domination. Certainly there is a point at which the influx of cash outpaces the cost of maintaining your own scientific research(except for the 32 turn nearly worthless tech that you'd research).

So for those of you with experience in the later game:
How much does purchasing a tech usually cost you? I've bought techs for only 100 gold in the Ancient times, but I have no clue how much something like Rocketry or Superconductor might cost. Certainly the number of Civs that have the tech are used by the AI to keep the costs even, as well as the value assigned to describe the importance the AI places on that tech overall. So does anyone have an economy such that elimination of science provides enough surplus cash to buy tech in the late game?

[David Weldon]

I have found the price of a tech in gold to be just a little more than 50% of the beakers that would have been required to research it. This would mean that if you want to maximize things, go with 0 science and buy the techs. The only problem here is that if you do the beaker thing, you aren't helping your competitors, whereas if you pay someone gold then you are. Of course you benefit from cheaper "research", but they also benefit from having your gold. If you can buy from a weaker civ, no problem, but I try my best to never buy from the leading AI civs.

On emperor, I find that by the late medieval age I am the science leader, and I keep myself there by using newly upgraded military units to beat down the next closest civs. As a result, I don't buy leading edge modern techs, but I'm quite confident that they can still be bought for less than they can be researched. Of course, the relative impact of giving a democratic competitor 8000 gold in the modern age is much greater than giving a despotic competitor 40 gold in the ancient age...

[jack_frost]

Couple things I've noticed about tech trading.

The AI does compensate for reduced research costs, or: the ai pays much more for a tech if your the only one who has it.

This is why its important when dealing to all the civs you see that you start with the one who has the most cash. If you are getting their full income, try the next one (You can tell if they go to 'never accept' suddenly). Always trade to the richest first, the differance especially in the later industrial/modern is -not trivial-.

I always sort of noticed, but I didn't realize how massive till I started taking notes on one game. The following are actual trades, emperor difficulty. Each trade max'd:

"510bc: finish road to france.
145 gold/world map for ivory.
buy currency from iroq -> 220/world map
terr, 110/2 a turn -> curr (jap)
terr, 30/3turn -> curr (rus)
terr, 100 -> curr (ger)
terr, 90 -> curr (fren)
terr, 60 -> curr (amer)
terr, 50 -> curr (aztec)"

"130ad: Unfortionately, I slacked off, and someone got chivalry. I trade for it. It costs me 180/11t, and my map from the Iroqs.
world, 240/2t (jap)
world 15t (german)
world, 40/10t (eng)
world 80/8t (french)
world, 40/7t (amer)
world, 50 (aztec)"

"1070ad: Dealing scientific method:

indian - 75t
germ - 55t
aztec - 50t
persians - 600"

Notice as they get lower and lower in order. The exceptions happen because I hit their income. As the prices go up so do the differances. I've also noticed (but I didn't test heavily) that the amount they're willing to trade for two techs at once is more then two seperatly.

For example: offer monarchy and republic to the aztecs, and they'll give you more then monarchy seperatly and immediatly after republic.

So the best possible deals require a two tech (on two differant paths) lead at the same time. Not uncommon if you focus on one path, and the second place techer goes down another. Just remember to trade with him first. And check for new tech constantly.

I need to disagree with you on the "if you want to maximize things, go with 0 science and buy the techs". The fact is this: it will be -much- more to your advantage with a 100% science rate, and selling your techs to everyone else. Unless you have the patiance to every turn check every civ to see if they have a new tech, buy it, and then sell it to everyone else (they usually wait a turn to sell to everyone after finding it [or, I think turn upkeep happens for everyone, and you just go first in every turn is actually how it works]). But I'd also rather control the pace of tech trades.

Also note.. it seems (this one is just a feeling) to me that AI tech climbing slows down when you deal heavily in tech. This makes some sense. If you are eating all your opponents money, they A) can't trade with eachother as effectivly, B) can't research themselves as effectivly, C) I think the AI does the same thing you do. If its the tech forerunner, it uses its income to increase research - and strongly keeps its lead. And in reverse, if its entire income is being used up, it'll slow down its research speed to compensate. Thus they become increasingly dependant, and you get more mighty.

Last tip: try to pay attention to when deals end. And time new deals to take advantage of the newly freed up income before another AI does. Specially important once everyone starts going broke due to huge debts to you.


Final note: I could be wrong I do not think buying tech is valid once you hit industrial. Although I usually live buying tech until education cancels out the library (which I never snag). The thing is, costs for tech grows in cost constantly, but jumps at every age. 10 gold might get you pottery. But nationalism will easily cost you in the thousands. Also, you will still need the science buildings for culture, so you can't really max out hard on economy / military, unless you just plan to raze every city you take.

I've been selling early modern techs in my current game for around 150g a turn. The cap at this point (at least for the first guy or two I sell to) is not what they are willing to pay - it is very much their income and what they have on hand. I also sometimes sell cities for less. If the computer is willing to pay 3000g for a tech, it'll cost you twice that to buy it. At that point renewing your average luxury costs around 1200.


**** that was a long rambling message... I should post to forums at 6:30am jacked up on caffine.

[Grumbold]

I think the cost may vary according to difficulty level or the attitude of the empire you trade with. I try and make sure I have one or more good military allies in the midgame and try to trade with them first to maintain good relations. I undersell the product slightly or overspend slightly to get what I want and keep relations optimal. Industrial era techs have rarely cost me more than 500 on that basis and can often be discounted further by offering them a world map (that hasn't changed since I did it 10 turns previously but still seems to carry a high market value). It's normally at this point that I can afford to buy a tech outright rather than trading one for one and helps ensure I hit key upgrades like infantry and tanks while my enemies are still playing with cavalry and riflemen.

[jack_frost]

I'd be interested to see how AI trading habits change depending on attutide, and how atitude is affected by trading practices.

From my experiance, unless you're just not dealing with them or doing something specific (getting too much of a lead, sitting units within their borders, etc) they will pretty much address you politely.

Gracious only happens when in a mutual protection for me, or right after a gift. How many shades of polite there are - beats the hell out of me.

This kind of thing is difficult to test.. maybe I will later.



In reply though, buying techs in late ages always seem prohibitivly expensive for me, unless I can immediatly trade it with a large number of other civs. Otherwise you start to struggle. Your always catching up. All your strength just ends up strengthening your opposition.

Also the cost of buying a tech shouldn't just be looked at as the on the table in cash cost. Its also lost trade revenue for finding it yourself.

If I can make a discovery every 4 turns. And sell it for a total of 100 gold a turn. Thats 10 a turn from 10 AI (really low end). I can end up with 500 gold a turn coming from other civs alone. In mid industrial you can pretty easily have +500 a turn consistantly if your aggressive.

But maybe its all really just a matter of play style. . . .

[inca911]

Jack_Frost:
So when looking at my two options, it sounds like you are an advocate of being the Science King. In my current game (the only one I've played), frequent warmongering to eliminate both the Indians and the Japanese has led to my being slightly behind in culture and about 4 techs back as well. However, I do have nearly complete control over a populated land mass equal in size to that of the 3 other civs combined and with a whole lot of extra luxuries/resources to boot. So do you have any advice on catching up? I found the English to be similarly behind so we traded 2 techs for 2 techs (I was behind by 6 advances!). My military juggernaut is transitioning to a peaceful cultural/scientific machine and I plan to transition to naval dominance to maintain the peace on my continent. The Japanese have 2 cities that will be destroyed very soon and the Germans have 2 cities on my continent that I plan on absorbing in due time. I do not have enough Gold to buy the techs I need, nor do I have trans-oceanic trade as an option yet. Any thoughts?

[David Weldon]

Jack:
Early in the game, even at 100% science, you can't come even close to keeping up with the AIs at higher levels. Thus you don't have the option of selling new tech to everyone else. That's why I advocate buying tech instead (and then selling it to whoever will buy it, of course).

Later in the game, once you have established a large empire size advantage to compensate for the AI bonuses, then you can switch to science instead of buying techs, and I agree with you that you can start to sell them at that point. I generally don't sell a tech until one of the AIs researches it by themselves, however. If I'm getting new techs in 4 turns, even if it takes the next closest AI only 5 turns to get a tech, that means I'm stretching my lead as time goes by. When you factor in the fact that AI often research marginally useful techs, you can really start to build a commanding tech lead. I got through almost the entire Industrial age while the AI were just getting to Industrialization this way. I especially protect Steam Power and military techs.

Inca:
Might I suggest a war of conquest? If you're winding down the military operations on your own continent, get a few transports and ferry all but a token force to the other continent. Raze a few cities quickly (you don't want to defend them), and then sue for peace and tech before the AI can counter-attack on your own continent. Treat the troops you send over as expendable and simply disband them or (if you're lucky and they survive) send them after another civ for more tech. The key is you have to do enough damage quickly enough to bring the AI to the peace table before he can sail troops to your continent. Use massive overkill!

This will have a side-benefit of disrupting their research efforts a little bit.

[regoarrarr]

Now, I know many have said that they have gotten advances in less than 4 turns, and I am one of those.

But now, I'm wondering if I didn't just see the report down in the bottom right corner saying Steel (3 turns) - or whatever the advance was, after I had already researched for a turn. In other words, it was 4 turns, but I just didn't notice it till 1 turn had already passed by.

Can someone check this out and post for sure on this? Because I was doing some more research and couldn't get the advances under 4 turns no matter what in some rudimentary testing.