Spend GS or save like always?

S.K. Ren

Prince
Joined
Apr 29, 2010
Messages
369
Location
Gig Harbor, WA
So I've been trying to figure out what's better between settling GS and saving them. And while it looks like settling early GS has more benefits, I'm seeing numbers that say otherwise. For the following examples I'll be assuming the following information based on my own experiences. The goal is sub 250 science win.

  • Each Academy gives +8 :c5science:, +10 after Scientific Method
  • For the sake of the experiment lets say you get your first GS (Excepting Babylon, I'll talk about them later) around turn 110 from Liberty finisher or turn 150 from an University with both the Garden and National Epic.
  • NC built by turn 80
  • University built on turn 130
  • Rationalism Opener taken on turn 150 (Unless you go out of your way to cripple your culture I've yet to pull this off save for one when I was within 6 turns of doing it)
  • Scientific Method is gained on turn 170
  • Free Thought policy taken on turn 180 (I usually get any Observatories out at this time as well)
  • Research Lab built on turn 230

So following those changes we can break up the gain of a single Academy into the following:

Turn 111 to 130
8:c5science: Base +50%(NC) *20= 240 :c5science:

Turn 131 to 150
8:c5science: Base +83%(NC+Uni) *20= 292.8 :c5science:

Turn 151 to 170
8:c5science: Base +98%(NC+Uni+Rat.Opener) *20= 316.8:c5science:

Turn 171 to 180
10:c5science: Base +98%(NC+Uni+Rat.Opener) *20= 396:c5science:

Turn 181 to 230
10:c5science: Base +115%(NC+Uni+Rat.Opener+FreeThought) *50= 1075:c5Science:

Total of 2320.6:c5science:

At this point you'll be up to 10:c5science: Base +165%(NC+Uni+Rat.Opener+FreeThot+R.Lab) = +26 bpt. Over 8 turns this only boosts GS Bulbs by 208:c5science: each.

Now when saving GS for a Science VC you want to bulb 8 turns after you've finished all your Research Labs and are running full science. On 4 city Tradition I get around 700 bpt w/o any academies. 700*8= 5600:c5science:... so why do I want to settle GS again?

As for Babylon. Settling their Writing GS gives them enough of a boost to land the Tradition straight into Rationalism timing without wasting a policy. Which ultimately pushes most of the times I used back by quite a few turns, meaning much more bpt, but my point still stands. Discuss?
 
Now when saving GS for a Science VC you want to bulb 8 turns after you've finished all your Research Labs and are running full science.

Hmm thanks for this! :goodjob:

I always burned all my gs immediately after building research labs...i wondered why i didn't get all beakers i was waiting for.

About OP :

For sub 250 turns games just keep them all. I often get my first gs around turn 135-140(i often get 2 ge first) and i dont see the point to settle it when i'm going to finish the game 100 turns later.
 
Rationalism opener applies on empire :c5science:, so

Turn 181 to 230
10 Base +115%(NC+Uni+Rat.Opener+FreeThought) *50= 1075 :c5science:
would be (10 Base +100%(NC+Uni+FreeThought)) + 15% Rat.Opener = 1150.

Since raw :c5science: boosts RA too, it produces more :c5science: than this calculation too.
(if you sign 10 RAs, it's at least 50turns of :c5science:)
 
Rationalism opener applies on empire :c5science:, so

Turn 181 to 230
10 Base +115%(NC+Uni+Rat.Opener+FreeThought) *50= 1075 :c5science:
would be (10 Base +100%(NC+Uni+FreeThought)) + 15% Rat.Opener = 1150.

Since raw :c5science: boosts RA too, it produces more :c5science: than this calculation too.
(if you sign 10 RAs, it's at least 50turns of :c5science:)

So the 15% from Rationalism is Multiplied After? That's even better, but my point still stands. Only Babylon can get a GS early enough that settling it provides a net benefit over saving it.
 
I think you are looking at this a little simplistically (no insult intended).

What about the net affect of attaining all of those policies X turns earlier, by this I mean that you are only calculating the affect on the ACADEMY, what about the affect on your empire science production, ie Research Labs x turns earlier mean more science. What about the advantages of the bulbing X turns earlier (ie after REsearch Labs) ? What about Public Schools, Universities (assuming your GS come from wonders/social policies/etc), obervatories.

On top of that, what about all of the other, non science oriented, net affects that are gained earlier; impossible to calculate but worth noting I think ?

I think your formulae is incomplete. You need to factor in the net science output of the empire as a whole (difficult as there are many variables). Perhaps start with an OCC example and work from there.

Thats just my analysis, apologies if it comes off condescending as that is not my intent. I do believe your insights pose interesting questions.

Let me know what you think, perhaps I am looking at this the wrong way :)
 
Well you're half right. See, yes settling Academies does let you get to those milestone techs quicker but the problem comes in figuring out just how much faster it really is. Unlike Babylon which is built for mass producing GS and gets access much earlier and faster, most people will only get 2-3 , after Acoustics - too late to help with the 1 Policy Branch into Rationalism timing - and right before Scientific Method in which case settling would only have 10-20 turns giving you about 300:c5science: and saving you about 1-2 turns... that not worth losing at least 2.5k in difference from bulbing at the end. In addition this only reflects the one out the longest, any other would be worth much much less. OCC and maybe 2 city is the only case in which I would advise settling simply due to your normal BPT being far too low to succeed at higher levels. In that case you will be relying on RAs to get the most out of your Academies. I mean if you somehow managed to get and keep obscene levels of gold for RA's then by all means settle and run RA's as often as you can. I imagine that will be more effective, a much quicker Plastics that's for sure, but my issue is with going for fast finishes using 4 city setups... maybe 5 cities would work better, given how Mercantile CS work.
 
Hmmm, I see what you saying.

QUESTION : How many GS are you considering for a non Babylon/Sweden Civ, prior to Research Labs and from what sources ?

I also think you need to add in RA benefit, which would also have to include the benfits of getting to science oriented buildings quicker. Again stipulating the number of RA's signed so people can compare with their own playstyle.

I really think you need a table with turns each science building is created, gs made/bought, policies/techs/wonders reached. I think the current example has too many missing variables to make definite conclusions regardless of the large apparent descrepency. You may well be 100% correct in your conclusion, I just think its still in the assumption stage at the moment. As mentioned previously I mean this only as a critique not an attack. :)

Personally I believe that you are likely correct, but I do not think the descrepency is near as large as it seems, not to mention the incalculables at play.
 
Agreed but like I said unless you have a windfall of gold for RAs, the gain will be negligible. The formula I've seen floating around for RAs is :

Floor[(Average BPT_Yours/2 + Average BPT_Theirs/2) *10, Average BPT_Yours *10]

Assuming this is true, and since we're looking at the space between Education and Scientific Method.

Academy
Value - 8:c5science:
Effective RA bonus: 40 - 80:c5science:

with NC
Value - 12:c5science:
Effective RA bonus: 60 - 120:c5science:

with NC and Uni
Value - 14.64:c5science:
Effective RA bonus: 73.2 - 146.4 :c5science:

with NC and Uni and Rat.Opener
Value - 16.84:c5science:
Effective RA bonus: 84.18 - 168.36 :c5science:


After Scientific Method it goes up to:

Academy
Value - 10:c5science:
Effective RA bonus: 50 - 100:c5science:

with NC
Value - 15:c5science:
Effective RA bonus: 75 - 150:c5science:

with NC and Uni
Value - 18.3:c5science:
Effective RA bonus: 91.5 - 183 :c5science:

with NC and Uni and Rat.Opener
Value - 21.05:c5science:
Effective RA bonus: 105.25 - 210.5 :c5science:


The only time you will get more than the minimum contribution is if your partner has a higher BPT than you Actually its backwards, the only time you will get less than the maximum contribution is if your partner has less bpt than you. The lower end represents your own contribution to the RA while the higher end represents the maximum you can get from your partner's contribution.

Those numbers are far too small to make up for the 3280:c5science: discrepancy between settling and bulbing at 700 bpt average, at least until you get PT and Scientific Revolution.

Speaking of which; with all the science buildings built and up through Free Thought taken, each citizen is worth:

Normal: 4.6:c5science:
w/NC or Obs: 5.75:c5science:
w/NC and Obs: 6.9:c5science:

With this you can easily calculate your endgame max science output. I used 700 as my sample bpt average because that my average, well... on average. Looking at the other end of the spectrum, in order for that first Academy to pay for itself simply by existing, you need to be doing as badly as 290 bpt when you bulb.

I still hold that settling will have almost no impact on how fast you get from Education to Scientific Method barring you screwing around before teching up. However I do acknowledge that the gap between Scientific Method and Plastics could use some shortening. bpt during that phase, for me at least, is around 250-500 bpt. So you would need 1-3 RAs to shave a turn off, I usually get 8 RAs off or so in this period so lets say its 5 low and 3 high for 6 turns saved... well given the difference of 200bpt before and after Science Labs, that's 1200 science gained from earlier labs plus the RA bonus at 2100 (half with PT, half with PT and Sci. Rev) is 3300. Close... better than close actually. So as long as its present for at least 8 RAs during the period between Scientific Method and Plastics in the optimal location (nix Obs since those aren't always guaranteed) it does better than paying for itself... hmm, so the cutoff for saving and planting should be right before you start your first wave of RAs during that period. Wow. glad I decided to write all this out.

Finally, an answer for your question: Around 12-13 GS, 2 (Hubble) +1(PT) +2(Faith) +7-8 Natural, most of the naturals pop waiting for Apollo
 
But what if those 2-3k earlier beakers was what allowed you to get, say cannons, in time to defend against an otherwise game-ending dogpile DoW?

The latest changes (GaK) have made C5 more than a 'going through the motions' *finally!* as there are now sufficiently many ways in which the AI can throw you a curveball .. so I'd say if your situation allows for it, ie. if you have enough protection be it military or geographical, and you are certain you can make it to the key wonders in time then it makes sense to save them .. otherwise I would settle or even bulb if pressed.
 
" Floor[(Average BPT_Yours/2 + Average BPT_Theirs/2) *10, Average BPT_Yours *10]"

-- not really, though I guess with some math manipulation it can seem close to that (though SUM(x) vs. AVG*x is not the same).

RA output = (SUM_bpt_yours/6+SUM_bpt_theirs/6)*multiplier

SUM = over RA length (30 for standard speed)
multiplier = usual 50%+25/25 (PT/SR).

So (when it works right), it's effectively 5-10 turns worth of beakers. GSs being a direct 8 turns.

minor side note is that you forgot the Academy boost at Atomic Theory which brings the base to 12 (post Plastics, so this is just an addition to already settled GSs not a reason to settle more).

edit: Effectively, you're best bet is to stop settling GSs around the time you're pushing to Plastics. Or ~500 bpt pre-Research Labs (assuming left side of Rat. is finished).
 
edit: Effectively, you're best bet is to stop settling GSs around the time you're pushing to Plastics. Or ~500 bpt pre-Research Labs (assuming left side of Rat. is finished).

I think this is way to late to be settling GS if you are going for a fast Science VC?
Not sure how settling at T150+ is going to be better than bulbing 70 turns or less later

I agree its more than just a simple beaker comparison as if settling gets you to wonders like Pisa and PT then you should do it but otherwise I'd need to be convinved your are better off settling more than 1 or 2 GS...
 
I would also add that your formulae for population and science does not include science output from trading posts and jungles. These are only relevant in the turns in which science buildings are obtained earlier from academy settling, but play into both RA's and raw science. This would also need to include the Academy boost Madjinn mentioned. NOTE : Obviously jungles are situational but a jungle city can make a big difference once the multipliers go through it, especially once trading posts start dropping science.

It is unclear to me exactly how many turns are being saved at each juncture via settling and therefor it makes it difficult to calculate through the benefits.

Your RA calculations were certainly interesting.

I think that when you start to factor in all of the smaller contributers ie trading posts, jungles, wonders, observatories (where applicable), population level variance etc they will end up eating away at the beaker deficit, especially once RA's are factored in. Each factor may seem somewhat unimportant but together they should make a bit of difference.

For me I would go back to saves (assuming you have them) at the important tipping points, record the bpt you were receiving to give an idea of how much you would get per turn by saving time. This would then factor in all of the relevent metrics affecting your empire at that time. Each time period needs to calculated seperately, by this I mean that savings between university and public schools, also yield savings between public schools and research labs. In other words a two turn saving, becomes a four turn saving (in effect) as each science building/wonder is being obtained two turns earlier. It looked like you were only considering the initial saving and not how it compounds. This applies to RA's as well. Also remember to add the rounds of science after bulbing as the academies will allow bulbing earlier and thus will receive all of the extra sceince at the end, between the two bulbing dates. Using your example this will be 700bpt. You do not need many turns to make up the descrepency, especially once the other savings are considered. Again using your eaxmple, and without adding in the earlier affects, you only need a 5 turn reduction to come out in front, and this is without your RA calculations EDITED : This is incorrect. Not sure what I was thinking here, had a brain-intterupted moment.

So any time something affecting science is built/researched/taken you need to calculate the affects on the saved turns. It is by no means a simple calculation and IMO cannot be approximated via the population metirc you employed. I think this lies at the heart of your discrepency.

In short I think Madjinns estimate is likely to be correct. :)
 
Personally, i find it hard to grow cities with GS's settled, especially if random gods take away Maritime CS's, or Austria consumes them along the way, or the map is just scarce on food. I know i can settle them on top of hammers, but that limits hammers too. I do settle one if i get it early (t.100-120 ish). Rest are saved (or burned for early scientific theory). Good enough for t.240 space Vic.
 
" Floor[(Average BPT_Yours/2 + Average BPT_Theirs/2) *10, Average BPT_Yours *10]"

-- not really, though I guess with some math manipulation it can seem close to that (though SUM(x) vs. AVG*x is not the same).

RA output = (SUM_bpt_yours/6+SUM_bpt_theirs/6)*multiplier

SUM = over RA length (30 for standard speed)
multiplier = usual 50%+25/25 (PT/SR).

So (when it works right), it's effectively 5-10 turns worth of beakers. GSs being a direct 8 turns.

minor side note is that you forgot the Academy boost at Atomic Theory which brings the base to 12 (post Plastics, so this is just an addition to already settled GSs not a reason to settle more).

edit: Effectively, you're best bet is to stop settling GSs around the time you're pushing to Plastics. Or ~500 bpt pre-Research Labs (assuming left side of Rat. is finished).

Well that formula is pretty much identical in the end.
Spoiler :

Ah, typo, the first expression needs to be *5 not *10 and I meant MIN not FLOOR
MIN[(Average BPT_Yours/2 + Average BPT_Theirs/2) *5, Average BPT_Yours *10]
MIN[(Average BPT_Yours + Average BPT_Theirs)/2 *5, Average BPT_Yours *10]

SUM_bpt over 30 turns

SUM_bpt /6 *0.5 = ((SUM_bpt /30) /2) *5

No real difference.

As for discounting Atomic Theory, by the time you get it you've got about 1 maybe 2 RA waves left in the game and you'll be running out of money (At least I know I am). The reason I'm focusing on the period before Plastics is because that's the last huge Science boost short of going Order for Planned Economy. And yes it does continue to benefit your RAs beyond that but the actual contribution per RA for each academy is small, even more so as you pull ahead of the AI in terms of BPT. It took 8 RAs to barely surpass the value of bulbing it and that was being generous on how many turns it shaved off getting Plastics. So settling anything beyond that is pointless, you won't make up for the lost beakers.

@Gabriel: I am not discussing the power of getting RAs or other sources of science. This is meant to discuss purely the value of an Academy vs Bulbing on an individual basis and how much benefit it provides in addition to normally signing RAs/Working jungles/specialists/etc. My population estimate was for the population only, not the tiles they work. You can easily tell how much science you get from those but the value of a citizen by itself isn't perfectly clear (though the pieces are all there).

edit: Doing a full write up. Will post later today.
 
i think the optimal play is generally going to be to plant the first few at most, then use the remaining to bulb. not necessarily save for the end to bulb, but bulbing critical techs earlier. things like getting to public schools 8 turns earlier will pay dividends, even though you're not immediately getting as many raw beakers as you would from a later bulb
 
i think the optimal play is generally going to be to plant the first few at most, then use the remaining to bulb. not necessarily save for the end to bulb, but bulbing critical techs earlier. things like getting to public schools 8 turns earlier will pay dividends, even though you're not immediately getting as many raw beakers as you would from a later bulb

possibly, ofc bulbing and instant buying Public schools or research labs will pay back in more GSs (or at least sooner).
 
Mase an error in my last post. Corrected it in strikethrough. Not sure how I came to the conclusion I did, I am going to blame the lateness of the hour :)

@ S.K. Yes I see what you are saying I just don't see how you can compare academies without looking at the extra turns that come from other sources due to academy planting (or early bulbing as Vexing suggested). Extra turns make m ore difference the higher the bpt , hence the tradiong post/jungle comment. As mentioned previously I feel RA's need to be considered as Academies play into them, in addittion to the effects of reaching the bulb date earlier via planting yielding extra bpt that carries through the calculation. You have motivated me to create a spreadsheet to examine the results and test some of the variables. :)
 
The difference is regardless of whether you plant the academy, you should be making RAs. You don't make RAs because you planted an Academy, rather you plant Academies because you plan on making a lot of RAs. That's what I meant. Regardless of whether I even have GS to spend, I will be making RAs up the wazoo and working TP jungle tiles.
 
S.K. Wow you respond fast :)

Yes I understand what you are saying, I don't think you are following what I am saying. (Again meant with no negative connotation.)

What I am saying is this. Let hypothetically suppose that plantng a GS saves you 1 turn, so each science milestone comes 1 turn ealrier. The net BPT you gain is going to be the EXTRA bpt on that turn that is gained as a result of the science buf that comes from the milestone ie Academy buff, Public School, Observatory etc.

The extra (delta) needs to be calculated at each successive milestone and then added together.

The greater your bpt (pop, jungles, trading posts etc) the greater the net benefit from planting as the delta you save will be bigger.

This in turn feeds into RA's as the RA value is dependent on your bpt (less or more dependent on you partners bpt).

So its not enough to say that you are working every science tile, and signing every RA if you are not factoring them into your delta savings ? For all I know a greater bpt may weaken my arguement as it will mean faster tech aquisition and therefor less academy effect. Then again it means any savings become magnified. I imagine the delta savings will end up reflecting an inverted hump, ie planting savings increase up to given bpt, then decrease as the academy contribution becomes less significant.

So I cannot see how an analysis can be correct if it does not factor in bpt at each given milestone ? I truly don't have definite conclusion of what I think the result will b e, just feel you can't draw a conclusion without examining all measureable factors.

Perhaps I am looking at this wrong ?
 
Haha, it seems we keep talking past each other. I get what you mean now. Yes, getting University/Labs up faster provides additional yield from all sources so I agree that should be noted.
 
Top Bottom