Need some clarification on the Research Agreement

[Mutineer]

I can add to what Browd said that science from RAs unlike from science buildings has no impact on amount of science from GSs bulbing.

Information for the people which care on the quickest possible victory and read this thread - there is no doubt that rush buy science buildings is better in your turns time finish.

Assuming tall empire. If you run wide empire ( my typical one consist of 8 min usually 10 cities) rush buying is waste of money. RA beat it down long way.

 

[fallout3dc]

I faced heated opposition on the evaluating my peaceful SV progress thread, but my ballpark math was not debunked. I will try to be a little more careful here. I am focused only on the payoff of beakers for gold, and I don't disagree that rush buying has other advantages besides beakers.

1. Estimate that rush buying a science building in a weak production gets you about 12 turns worth of extra beakers in that city.
2. Assume you have four cities, and this city is a quarter of your science, so that's roughly 3 turns of empire science.
3. For the cost of a science building, you could afford at least 2 RAs.
4. Value each RA as shaving two turns off research time, so each RA is worth 2 turns of empire science.
5. 2 RAs x 2 turns of empire science == 4 turns of empire science.
6. 4 turns of empire science > 3 turns of empire science.
7. qed: RAs have better gold:beaker value than rush buying science buildings.

This approach is very rough but it scales for era, and I think is conservative on all points, except that (4) RA payout is closer to 1.5 turns of empire science (and RA still breaks even with rush buying science buildings). OTOH people would prolly still rush buy science buildings if they were twice as expensive, so RA still a good deal. I have numbered the points so you can tell me where my logic breaks down!

Yeah, you and I did a good job in helping the poster with his science victory. I wish he actually ended up replying to that thread, so I could keep helping him, but we kinda hijacked that thread for research agreement debates.

 

[beetle]

there is no doubt that rush buy science buildings is better in your turns time finish

Absolutely, this point is not in debate. I am asserting that, based on the rough math, RAs are even better!

before you continue to go on about how much better RA's are, shouldn't you have BNW first, so you at least have some first hand experience with the new RA system?

Lucky for you that I am still on GnK or I prolly never would have noticed that the conventional wisdom that RAs are weak doesn't hold up to scrutiny. (People were saying that back in the Gnk days too.) Asside from RAs being half as good as they were, what else might I be missing? An RA being worth ~2 turns of empire science is the BNW value.

Just as gold now is more valuable than gold later, early beakers are more valuable than later beakers

Absolutely. But I think the metric of “turns of empire science” addresses this. So much easier than estimating beakers, then comparing early beakers to late beakers. I will confess some lingering concern that this metric may be obfuscating rather than finessing the 30-turn delay aspect of RAs, but I think it's legit.

In your example, you assume the early science building will take 12 turns to hard-build, but that rush-buying only accelerates your next tech by 3 turns. That may be an OK assumption in general terms, except it misses the point that the early science building will accelerate every subsequent tech as well.

If Technique A outperforms Technique B, and Technique A is available earlier in the game than Technique B, logically it has to follow that Technique A leads to better acceleration/snowballing than Technique B.

Also, the extra science building may, at the margin, yield you one more (or at least an earlier) GS, which the RA will not do.

Good point, as is your observation that RAs can be lost to a DoW (or an AI being eliminated). Also the observation that RAs benefit your opponents. Also, needing DoF might limit your preferred play style. Still, no one has challenged my rough math that, in terms of beakers bought by gold, RAs are competitive with rush buying science buildings.

The more common choice, in my experience, is either (1) doing the RAs now and deferring rush-buying the science building until I have the gold

If my math is right, delaying rush-buying for RAs is a net gain of beakers.

or (2) rush-buying the science building now and doing one RA a few turns later and another some number of turns after that

If my math is right, passing on an RA opportunity so you have enough gold to rush buy ASAP is a net loss of beakers.

There was some concern [from the other thread] that a player could never get enough RAs going to make a difference. A fair benchmark for rush buying after NC is 3 universities and 3 public schools, hard building in the cap only. That would be enough gold for 12 RAs during the course of game, but if RAs out perform rush buying at 4:3, one would only need 9 RAs to test this. With 7 civs on a standard map, 9 RAs seems pretty feasible for the course of a game. Optimal would be a mixed approach: aggressive RAs with rush buying in lowest production cities, and rush buying in other cities only to the extent that RAs are not passed over for lack of gold.

 

[Browd]

You keep saying "if my math is right," but it isn't right. You are ignoring (actually you are cavalierly dismissing) the time value of when beakers are delivered and not accounting for risk of loss of the entire RA (notwithstanding your acknowledgement of that risk, your math applies it a zero risk-weight).

To make my time-value point clear, if RAs didn't pay off for 150 turns, but still delivered, on turn T+150, the same 1-2 turns of science, would you still view that as a good use of today's gold vs. buying a science building today? Shouldn't your math include a discount rate for future beakers vs. today's beakers? Shouldn't your math apply a risk-weight to the expected yield of the RA (given the risk of cancellation from a DOW or denunciation by your RA partner or your RA partner getting conquered)?

 

[beetle]

You keep saying "if my math is right," but it isn't right. You are ignoring (actually you are cavalierly dismissing) the time value of when beakers are delivered and not accounting for risk of loss of the entire RA (notwithstanding your acknowledgement of that risk, your math applies it a zero risk-weight).

I don't think it's fair to characterize me as ignoring or cavalierly dismissing the time value when I have been actively trying to put parameters around it.

It is a very fair point about the risk of loss of the entire RA. I did not think of that earlier. I would estimate it at least 25%, which weakens RA to the point where any gold:beaker advantage over rush-buying disappears.

In BNW, do RAs get canceled just because a DoF expires? (Not a DOW, just the DoF being interrupted?)

To make my time-value point clear, if RAs didn't pay off for 150 turns, but still delivered, on turn T+150, the same 1-2 turns of science, would you still view that as a good use of today's gold vs. buying a science building today?

Excellent way to challenge my “empire turns of science” metric! Thanks for that!

It feels so wrong, but my first impression is that the time RAs take to payout ends up being irrelevant! (How can that be?) I will think about this more, and maybe in the meantime you or someone else can figure it out.

If at turn 100 you could spend 250 gold to shave two turns off of whatever you were researching on turn 250, would you not jump on that?

 

[kaspergm]

In BNW, do RAs get canceled just because a DoF expires? (Not a DOW, just the DoF being interrupted?)

Don't you get a reduced yield if the DoF is no longer active by the time the RA expires? Or was that G&K?

 

[IronfighterXXX]

Absolutely, this point is not in debate. I am asserting that, based on the rough math, RAs are even better!

Enough with this empty talk - show us in game that RAs are better!

There was Korea, immortal, science GOTM lately http://forums.civfanatics.com/showthread.php?t=528876

I finished on 217 T without even one RA, however bought ~ 6-7 science buildings during the game - beat this and show us RAs are better in turn times finish!

 

[Browd]

If at turn 100 you could spend 250 gold to shave two turns off of whatever you were researching on turn 250, would you not jump on that?

I certainly would not. I suspect I would have far more valuable uses for that 250 gold now, whether befriending a CS or upgrading 3 archers to Comp Bows or whatever.

Don't you get a reduced yield if the DoF is no longer active by the time the RA expires? Or was that G&K?

There is no diminution in RA payoff due to DOF expiration at the end of its term.

 

[beetle]

...given the risk of cancellation from a...denunciation by your RA partner

In BNW, does a denunciation by your RA partner (presumably during the DoF, so a backstab) cancel the RA? There is no such effect in GnK.

I certainly would not. I suspect I would have far more valuable uses for that 250 gold now, whether befriending a CS or upgrading 3 archers to Comp Bows or whatever.

Agree, the opportunity cost is high that early in the game, and that distracts from core question as to the value of gold for delayed beakers. For the sake of argument, lets assume a game goes to turn 350: Would you pay 250 gold at turn 150 for 2 turns off whatever you are researching at turn 300?

Enough with this empty talk - show us in game that RAs are better!

I am not competent to do so. That does not invalidate the math.

I finished on 217 T without even one RA, however bought ~ 6-7 science buildings during the game - beat this and show us RAs are better in turn times finish!

If I could (and I cannot) that would only prove that I am a better player than you! It would demonstrate nothing about the efficacy of RA versus rush builds. To make such a test, one would need to change as few variables as possible. Differences in player skill is a huge variable, prolly even bigger than dirt! You might be able to try this yourself. One confounding variable might be the need for DoF. So you would need to play both games trying to maintain similar relations. With your example, do you think you could have completed 9-10 RAs by turn 217? Shorter games make it much harder to get enough RAs in to add up.

 

[Browd]

In BNW, does a denunciation by your RA partner (presumably during the DoF, so a backstab) cancel the RA? There is no such effect in GnK.

No, sorry, I forgot they patched the bug where denunciations caused RAs to fizzle. Only a DOW (or disappearance of your RA partner) will cancel the RA in both G&K and BNW.

Agree, the opportunity cost is high that early in the game, and that distracts from core question as to the value of gold for delayed beakers. For the sake of argument, lets assume a game goes to turn 350: Would you pay 250 gold at turn 150 for 2 turns off whatever you are researching at turn 300?

Candidly, no.

 

[Uzael36]

You keep saying "if my math is right," but it isn't right. You are ignoring (actually you are cavalierly dismissing) the time value of when beakers are delivered and not accounting for risk of loss of the entire RA (notwithstanding your acknowledgement of that risk, your math applies it a zero risk-weight).

To make my time-value point clear, if RAs didn't pay off for 150 turns, but still delivered, on turn T+150, the same 1-2 turns of science, would you still view that as a good use of today's gold vs. buying a science building today? Shouldn't your math include a discount rate for future beakers vs. today's beakers? Shouldn't your math apply a risk-weight to the expected yield of the RA (given the risk of cancellation from a DOW or denunciation by your RA partner or your RA partner getting conquered)?

You also forgot about the free hammers gained when buying the buildings as well. Now for those ~10-20 turns normally spent building the science buildings, you can build something else useful.

Anyways, the whole argument is moot. He has never played BNW to even backup his claims.

 

[beetle]

You also forgot about the free hammers gained when buying the buildings as well.

Actually, I mentioned that right up front on this thread.

Anyways, the whole argument is moot. He has never played BNW to even backup his claims.

Don't need to play to do the math.

 

[Uzael36]

Don't need to play to do the math.

You can theorize, but you cannot test your theories. They remain theories until you can test them. Your "math" is missing tons of stuff that all play a big factor in the end result. Until you can prove your theory covering all aspects of the game, or test it out with game samples, you have nothing but a theory.

 

[Mutineer]

And you just completely ignoring opposite point of view. Please tell me How Rush buying better for wide empires, like 10 city empire.
How increase in 1% research (some times even less if you ever used interfaith dialog for example),
better then 2 RA which benefit from constant fast grow of your wide empire?

 

[Uzael36]

And you just completely ignoring opposite point of view. Please tell me How Rush buying better for wide empires, like 10 city empire.
How increase in 1% research (some times even less if you ever used interfaith dialog for example),
better then 2 RA which benefit from constant fast grow of your wide empire?

My main point is that beetle is arguing without experience or enough math to back up his claim. I'd be willing to debate with someone who does play BNW and if good skill.

2 RA's from a wide or tall point of view gives the same result, which is what ever your research partner achieves, as it is based on the slowest party in the agreement. Your cities play no part.

But I suppose you are referring so how difficulty/costly it is to buy for all those cities. I honestly do not play wide. Not many do in BNW, and not usually for speed victories, but when done it is usually for ICS culture or domination. In that way, I do not know what would be better. I imagine you'd have other things more important than both. You might still buy in your capital, if you if you even care about science.

 

[Mutineer]

Ofcouse you care about science, but you just achieve it differently. Because your cities are generally MACH smaller then for tall empires you consume happiness at mach faster rate. If you get happiness you just focus your cities on grow and in a few turns all happiness is consumed. Then you switch back on production or specialists.

Basically wide empires have grow in time of:
1) Getting religious happiness. In ideal case you can achieve +5.5 happiness per city in extriame cases you can get +6 happiness from just your religion per city.
2) Getting Policy happiness, mostly applicable to wide coastal empire form exploration happiness +3 per city) some times from patronage or mercantilism happiness.
3) Getting ideology happiness.

So, you care about tech as a way to get to there milestones, After ideology happiness you will surpass any tall empires in everything fast.

And about RA, you never have gold to rush anything, you need all gold maintains or buy happiness by the way of city states. If you get some spare gold you will spend it on RA, as early on often you will be lowest in sci on start of agreement. If you ever used interfaith dialog you know that only 1/2 or 2/3 of your research coming form cities, other come form faith.
Rush buying effect DOes depends on your cities, as when rush buying in tall empires increase your sci for 1/4 of your pop in average, when for wide 1/10 in average.

 

[Acken]

And you just completely ignoring opposite point of view. Please tell me How Rush buying better for wide empires, like 10 city empire.
How increase in 1% research (some times even less if you ever used interfaith dialog for example),
better then 2 RA which benefit from constant fast grow of your wide empire?

There's no doubt 3RA are better than rush buying a univ in a crap 7pop city.

To me the biggest advantage of rush buys are the saved hammers allowing you to get other stuff. And we all know the faster your finish time is the harder it is to place all the buildings you'd like.
Saying that RA are better science per gold when ignoring additional benefits is one thing, claiming its therefore the best strategy while ignoring everything else is another. A good game isn't just about your bpt, you cannot ignore additional benefits when it comes to really play the game.

Here are some of the things rush buys allow:
-Faster GSs, may earn you that extra to finish 10 turns earlier.
-Frees up hammers for more buildings, so you get something else out of the rush compared to RAs. A wonder, a market, etc.
-Can help you reach techs you need right now, faster. The science is instantaneous rather than delayed. Important for reaching Renaissance for example.

These have to be considered.

And then it's not mutually exclusive, maybe the optimal strategy is rush buy in a city and take 3 RA rather than rush buying 2 or make 6RA.

 

[reddishrecue]

If you make some good food per turn in a city and you have science buildings such as academies and universities etc,, you can obtain a good science per turn and won't need to rely so much in research agreements. AIs also often like to ask more from you if you have a tech lead in science.

 

[fallout3dc]

Here we go again.....

 

[Paszczak]

For the sake of argument, lets assume a game goes to turn 350: Would you pay 250 gold at turn 150 for 2 turns off whatever you are researching at turn 300?

How many times does it have to said? Free stuff right now is more valuable than more free stuff much later. 30 turns (not to mention 150!) is a lot of time in Civ 5, you can advance 2 Eras and double your science output in that time.

Ask a Domination player - would he prefer to receive a free Composite Bowman (75) on turn 50 or a free Great War Bomber (325) on turn 200. The bomber costs 4 times more so it's clearly a winner?

Also, it's a bit naive to assume that a RA signed on T150 would shave 2 turns on an Information Era tech 150 turns later. That would require both you and your RA partner to produce at least 500/turn starting from turn 150. I know only one player that can do it and I know for sure that no AI in this game can do it.

Don't need to play to do the math.

Well in that case, I state it as a fact that teching to Machinery first, rush-buying Crossbows to kill your neighbour and Annex his Cap with University leads to faster winning time than spending that gold on your own University OR Research Agreements!

Now prove me wrong.