June 2015 Academy vs Bulbing Discussion

[GerrardCapashen]

This appears to result in if you have are playing wide, your costs are heavier for saving scientists than for peaceful four city tall. (And if you are conquering, your costs are higher yet)

This results in a highly player specific style as to which one minimizes your costs. But those who are quickest at winning (via having ruthlessly streamlined their play to avoid distractions) have less opportunity costs for hoarding GS.

I don't warmonger all that much, but I have found that once your conquest actually starts to make progress, having a large puppet empire with city connections to your capital gives you a lot of GPT. So my warmonger wide games usually end up with much higher GPT than my peaceful turtle games, and you could better afford the maintenance costs for your 8-9 GS or however many you are saving. Even still, I haven't noticed it to be a problem. Just send a couple external trade routes if you are really having money troubles.

 

[joncnunn]

I don't warmonger all that much, but I have found that once your conquest actually starts to make progress, having a large puppet empire with city connections to your capital gives you a lot of GPT. So my warmonger wide games usually end up with much higher GPT than my peaceful turtle games, and you could better afford the maintenance costs for your 8-9 GS or however many you are saving. Even still, I haven't noticed it to be a problem. Just send a couple external trade routes if you are really having money troubles.

It's true that puppets give more gold per turn (due to being locked on gold focus and prioritizing no maintenance cost buildings) than the increased unit costs.

But your unit maintenance cost for 8-9 GS was still more than it was for 8-9 of them for a peaceful 4 city player, and it's the unit maintenance cost (& turn number you win on) that determines optimum last turn for academy and not your net income per turn.

 

[GerrardCapashen]

It's true that puppets give more gold per turn (due to being locked on gold focus and prioritizing no maintenance cost buildings) than the increased unit costs.

But your unit maintenance cost for 8-9 GS was still more than it was for 8-9 of them for a peaceful 4 city player, and it's the unit maintenance cost (& turn number you win on) that determines optimum last turn for academy and not your net income per turn.

In the strictest sense that is true. However, I am going to argue that unit maintenance is so low relative to the value of a bulb that it is irrelevant and thus maintenance costs should never cause you to decide whether or not you should bulb or plant a GS.

I'm using the unit maintenance costs found at the bottom of this page:

http://civilization.wikia.com/wiki/Mathematics_of_Civilization_V


Let's assume the following hypothetical example:

You have generated 4 great scientists by turn 150, and 8 by turn 200. You save your GSes until turn 200, when you start mass bulbing them for whatever victory condition you are going for (Internet, Globalization, Hubble/science, or Radar for bombers).

In this hypothetical example, you have 30 military units and 10 workers in addition to your GSes.

On turn 150, you have 44 units, so you are paying 89 gpt for unit maintenance. If you planted them all for academies, you would have 40 units, and pay 81 gpt for maintenance. (89-81)/4 = 8/4 = 2. So each GS is costing you 2 gpt at turn 150.

On turn 200, you have 48 units, so you are paying 131 gpt for unit maintenance. If you planted them all for academies, you would have 40 units, and pay 108 gpt for maintenance. (131-108)/8 = 23/8 is basically 3. So each GS is costing you 3 gpt at turn 200.

If we assume that over the turns we are discussing the GSes cost an average of about 2.5 gpt, and you had on average 6 GS, then from turn 150 to turn 200 your GS costs were 50 turns * 6 units * 2.5 gpt/unit = 750 gold. Considering that 1000/1500/2500 faith is considered a reasonable price to pay per person for great people that you need, 750 gold total cost for 8 great scientists is outrageously low. Also, unless you are going Domination, you probably have a smaller army and fewer workers, so your unit costs per GS will be even less.


In addition to gold maintenance costs being irrelevant, I personally feel that the loss of 1 food or production due to planting is also basically irrelevant, as long as you don't block a Civil Service farm. The only things you should consider are: what turn do I expect to win, how many late game bulbs do I need, will I be going Freedom (+4 science obviously makes academies more worthwhile), does the NC city have an observatory, and do I need midgame science to get Renaissance/Industrial wonders or military units more quickly.

I went from planting 3-4 GS per game to planting zero, and I think my games have improved significantly. I might go back to planting one, but I think more than one is probably a mistake.

 

[Browd]

The point about the gold maintenance cost of hoarding GSs for late game bulbs was not that it is uneconomic -- in fact, as you note, the gold cost is quite affordable. It was to rebut the implication that hoarding a GS was somehow a costless alternative to giving up 1 hammer for an academy. There is a modest cost (gold or hammer) either way you go.

 

[GerrardCapashen]

The point about the gold maintenance cost of hoarding GSs for late game bulbs was not that it is uneconomic -- in fact, as you note, the gold cost is quite affordable. It was to rebut the implication that hoarding a GS was somehow a costless alternative to giving up 1 hammer for an academy. There is a modest cost (gold or hammer) either way you go.

I know. I am arguing that both costs (1 food/hammer or 2 gpt) are so negligible as to be not even worth considering from a strategic perspective.

 

[Browd]

I would agree with that.

 

[Acken]

Sure so to go back on topic the overwhelming evidence says to plant an early academy (or 2 if Babylon or Maya) otherwise save your Scientists to bulb after you have Research Labs in every city. The possible exception is to bulb scientists early to speed up Scientific Theory or Plastics providing you have the gold or means to build the Public School or Research Lab immediately.

But what about Culture Victories. How do you optimise science for this approach?

I think it would be good to have some more culture victory civs in the Diety game maybe France so this can be fleshed out a bit more.

Culture victories have an oversized production demand in the mid game (when you have to make schools, museums and archeologists and wonders and hotels). Speeding up toward that point when you have nothing to do and after that point to reach internet would be the best. You just don't want to bulb during that period.

The exception is Futurism CV. If that's your goal you should speed up toward ideology and hotels.

 

[joncnunn]

I know. I am arguing that both costs (1 food/hammer or 2 gpt) are so negligible as to be not even worth considering from a strategic perspective.

I would agree with that.

I am mostly theory crafting on which of the two opportunity costs would be smaller.
(Not that it really matters win you can win the game so fast with optimum play in Civ V.)

 

[Redaxe]

Question for Acken - When do you build your granaries? Before or after settlers? And do you change your priority if you have a heavy wheat/deer/banana start?
Does a granary increase your settler production if you have say 2 wheat so it's plus 4 food which I believe does have some value towards building settlers.

 

[TheMeInTeam]

When you mathematical guys have resolved the issue, could you please tell us in simple language whether we should be bulbing or planting the first natural one? I appreciate the number crunching you're doing. You do us a service, sirs.

I do too, but it's dangerous to take numbers at face value. They can be used well or poorly.

For example, consider the player chooses to bulb earlier (for less beakers) rather than later, but in doing so gets universities or public schools earlier by x turns. If the analysis factors in raw beakers from settling vs bulb early vs bulb late and makes a conclusion on that alone, it fails. You have to factor the opportunity cost of x turns of science buildings, and the gold cost of carrying a unit for dozens of turns. Those things don't necessarily offer enough weight to change the conclusion, but they're significant enough that in a rigorous analysis they should be factored.

 

[Acken]

Question for Acken - When do you build your granaries? Before or after settlers? And do you change your priority if you have a heavy wheat/deer/banana start?
Does a granary increase your settler production if you have say 2 wheat so it's plus 4 food which I believe does have some value towards building settlers.

After. Yes if there are no hills/mines and many granary ressources. Otherwise no, the conversion is bad.
Only food SURPLUS increases settler production. So you first have to "spend" some citizens getting into a food surplus. Then here are the ratios:
+1 food: +1 from food surplus
+2-3 food: +2 from food surplus
+4-7 food: +3 from food surplus
+8-11 food: +4 from food surplus

If we compare 2 farmed wheat + 1 mine + granary vs 3 mine:
1) 3P + Value_food(2+2+2*5 - 6) = 3 + 4 = 7 Production
2) 3P + 3P + 3P = 9 Production
If strategy 1 didn't have a granary:
1) 3P + Value_food(2+2*3 - 6) = 3 + 2 = 5 Production
If no tile improvements:
1) 2P + Value_food(2+2+2*4 - 6) = 2 + 3 = 5 Production
2) 2P + 2P + 2P = 6 Production

You add of course the city tile and palace production to these numbers. And then remember that every turn a settler is delayed is that many turns of production/growth lost on the expansion. And Collective Rule do not work with food bonuses.

I do too, but it's dangerous to take numbers at face value. They can be used well or poorly.

For example, consider the player chooses to bulb earlier (for less beakers) rather than later, but in doing so gets universities or public schools earlier by x turns. If the analysis factors in raw beakers from settling vs bulb early vs bulb late and makes a conclusion on that alone, it fails. You have to factor the opportunity cost of x turns of science buildings, and the gold cost of carrying a unit for dozens of turns. Those things don't necessarily offer enough weight to change the conclusion, but they're significant enough that in a rigorous analysis they should be factored.

We do factor that.

 

[TheMeInTeam]

We do factor that.

YOU factor that, and others in this thread did also. You also implicitly factor a number of things that don't come to mind easily, that change decisions on some turns, and you do it better than most. Those are the things that are hardest to replicate, and part of the reason that someone making similar large-scale conclusions to you still winds up 20 turns slower on the same start.

I don't know what all of those things are. I don't have thousands of hours in this game like I have in EU IV for example, but I know how this type of divergence happens at least. It was just intended as a caution; numbers are very useful, but using them only reaches so far as the factors including them.

 

[Acken]

What I meant is that the benefit of getting faster schools or labs from a bulb is factored in the discussion. In order to make comparison fair. And that if someone is waiting conclusions he should rest assured that this is considered.

 

[Acken]

By the way here is a spreadsheet that I have been working on since the creation of the thread that may help people decide. It is not bug proof

Manual changes and assumptions are in the green areas.

Quick conclusions:
First academy is better than bulbing scientific theory
First academy is worse than very late game bulbs and approximately equivalent to bulbing plastic
Each subsequent academy is worse and value decreases with game length.

 

[Acken]

Noticed a bug in previous version:

This version should now work properly. Note that some bug still remains if you put unlikely inputs like an academy after scientific theory or an academy before education.
I'm a bit lazy to make it foolproof to that kind of inputs, the spreadsheet should work for your reasonable situations.

http://www.filedropper.com/showdownload.php/academy_1 (click download, enter antibot code)

Nevermind the grammatical error "as to be" -.-

 

[fallout3dc]

Hold on, I can't keep up with this scientific mumbo-jumbo.

 

[Olodune]

Quick conclusions:
First academy is better than bulbing scientific theory

That is what I am seeing in your spreadsheet for any reasonable set of assumptions. The one I am least sure about is the beaker value of 1 turn early Schools. Currently it is set at a reasonable 250. But even at a very high 400 the early Academy still comes out ahead most of the time. Without New Deal the two options are very close for a fast game, and I haven't been choosing that tenet.

Personally, I am still surprised at the number of beakers coming through the Academy via bulbs. Makes sense though.

Despite all this evidence I still think it is usually correct to bulb ST if I am planning on hard building those schools. WF timing makes a difference, but that era usually has many hammer priorities and I still need to grow then.

 

[beetle]

Each subsequent academy is worse and value decreases with game length.

Could you clarify here? Does not the value of an academy increase the longer the game goes on? Or do you mean that each subsequent academy is worse and the value decreases the later in the game the GS is planted?

 

[tommynt]

there is no equation for when to bulb and when not.
Bulbs for example only help if u can benefit right away the bulb (f.e. buy/build Public schools) - but very often you want to finish a previous build or dont have money to buy the scientific building.
And then everything changes if u dont even go forspace win but want to bulb for example arties or nukes.

There are obviously some rule of thumb like - not building academies after like 50 turns before your planed finish time.

For me its pretty simple with academies - plant them allways if >90t till finish and never <50t till finish - think about in between

Whats game to game (situation) dependent is if using GS to bulb key techs is better as saving all till end. I usually bulb one or 2 and save rest.

 

[Acken]

Could you clarify here? Does not the value of an academy increase the longer the game goes on? Or do you mean that each subsequent academy is worse and the value decreases the later in the game the GS is planted?

Yes I meant that it decreases when game length decreases (for obvious reasons). And that each subsequent gives a lower value (because it comes later AND you get less extra bulbs).

there is no equation for when to bulb and when not.
Bulbs for example only help if u can benefit right away the bulb (f.e. buy/build Public schools) - but very often you want to finish a previous build or dont have money to buy the scientific building.
And then everything changes if u dont even go forspace win but want to bulb for example arties or nukes.

There are obviously some rule of thumb like - not building academies after like 50 turns before your planed finish time.

For me its pretty simple with academies - plant them allways if >90t till finish and never <50t till finish - think about in between

Whats game to game (situation) dependent is if using GS to bulb key techs is better as saving all till end. I usually bulb one or 2 and save rest.

There is an equation to tell you how much beakers you're likely to get out of it. Other considerations than the most beakers you can get is up to the player. How much extra shools would give the player has to be put manually by players.