Post-Patch Research Agreements?

[GiantCornSnake]

Hey guys, I'm having a hard time understanding how the research agreements work now in the 332 patch. I sign them but I don't know what they do now. I'm certain its really simple but I have yet to notice any change in my beaker count when I have research agreements going.

 

[Furycrab]

You'll basically get a bunch of bechers when the RA ends, the amount is supposed to be 50% of the median value of all available techs (Median, not average, so not researching naval techs won't really change much)

It seems to dump into whatever tech you are currently researching, but the overflow is really random, but it does overflow by just putting any extra bechers into a random other tech.

Not 100% sure yet. My impression of RAs is that they are more practical in that you don't waste a ton of time with blocked techs. However they are a bit weaker unless you get both the Tower and Rationalism. At that point they actually feel a bit stronger.

 

[ghoultek]

Fury,

I know the patch notes say median, but the in-game Civilopedia says its half the average costs of available techs. Did 2K issue a statement about the discrepancy or is it a situation where when in doubt trust the patch notes and ignore the in-game data source?

 

[Martin Alvito]

The median is a type of average. In common English usage, we tend to assume that the unmodified "average" should be interpreted as "mean", but that isn't necessarily always the case.

Any overage from a Research Agreement is applied randomly to an available tech. How that overage is assigned is not yet clear to me. I have had it assign the overage to techs that were largely completed, so it's not locking out techs using the 25% + 3 mechanism it was using previously.

RAs are considerably stronger than they used to be if you take Rationalism and build the Porcelain Tower. You can easily make four of the new RAs do the work of five of the old ones by manipulating your tech path to increase the median value of available techs, and four can even do the work of six at some points of the tree.

Couple that with the increase in late-game tech prices, and the result is that Rationalism is the single most powerful policy in the game for peaceful play, and the Porcelain Tower is the strongest Wonder for that style. The Hanging Gardens is debatably stronger for a pure warmonger, since it can easily be translated to numerous additional early units that become fully promoted.

 

[dubrown]

With what I've learned so far with the new RAs is that it isn't optimal with early RAs, sure, I'm only a third through my first game in the new patch but the RA I've signed so far didn't give me many beakers at all, most likely due to the fact that that I'd beelined up to chivalry I believe it was, all other expensive techs blocked by cheaper techs not researched yet, thus the net result of the RA was the median (or average) of a bunch of very cheap tech (2-~5 turns) and the expensive one (chivalry) I did research at that point.

So the lesson I learned from this, when signing RAs make sure to finish the cheap techs before the RA runs out and switch to the expensive one you want when the RA finishes.

Ofc, I should have thought about that before as it's all in the math, no matter if it's mean or median calculation. From my result though I would guess that it do calculate the median (which disregards of the lowest and the highest value when calculating the average) but it's just a guess, I haven't done the math yet.

 

[killerloop]

I also experienced that it was better to sign 3 RA's in 3 consecutive turns than signing them on 1 turn.

In my caseI was researching steel, and when the 3RA's expired on the same turn I got Steel and a little overflow. When I got the RA's on three turns, 2x beakers went into steel, and the last I redirected to compass. Got me more beakers, although in both cases I did have the same "available techs" that went in the calculation. (was on quick)

 

[ghoultek]

Thanks gents. I was wondering if there was any further clarification on how the feature was implemented.

 

[WeaselSlapper]

Ofc, I should have thought about that before as it's all in the math, no matter if it's mean or median calculation. From my result though I would guess that it do calculate the median (which disregards of the lowest and the highest value when calculating the average) but it's just a guess, I haven't done the math yet.

Mean and Median calculations are actually very different and your definition of median is not accurate.

A mean is the average of a set of values, where the median in the middle value. This difference can result in vary different results depending on what the numbers look like. Here's to examples:

Example 1
An early game RA (my numbers are not reflective at all of any beaker counts in the game just made of random numbers that seem to make sense). I have one high level tech researched and a whole bunch of lower ones left to research (I beelined some tech, and the tech i beelined opened up two higher techs)
100, 100, 100, 100, 100, 500, 500. The mean/average of this set is 214.28, while the median is only 100. There are some cases where people take off the highest and lowest values when calculating the mean, but that's not the case when choosing the median.

Example 2
A later game RA where I haven't researched any navel techs. Beaker counts of my techs:
100, 1000, 1000, 1000, 1000, 1500, 1500. Mean = 1014, Median = 1000.

Bottom line is beelining is now (assuming it does use the median and not the mean) very bad for research agreements while generally balanced research even if you have a few very low techs still available is the best way to maximize your return from RAs.

 

[formivore]

How is the median calculated if you have an even number of techs open? Is it the mean of your two middle techs, or something else?

 

[grandad1982]

I wish we know for certain how the overflow is distributed.

I had two RA come in about 10 turns (can't really remember) apart and both times the overflow went into Astronomy. I had 2 or 3 techs of the same price unlocked.

On another note I have been finding civwillard very useful for letting me how long till an RA comes in so that I can get more benefit from it.

 

[Randall Turner]

the mean of your two middle techs

this

I wish we know for certain how the overflow is distributed.

I had the same experience you did, ie, two RA's completing at the same time resulting in my currently researching tech being completed and then some additional random tech being completed - then I was able to assign the "overflow" however I wished.

It seems to pick one at random. In my case, it completed an additional tech that wasn't even available at the instant the RA's matured. (It became available when my original tech completed.)

 

[DaveMcW]

RAs are considerably stronger than they used to be if you take Rationalism and build the Porcelain Tower. You can easily make four of the new RAs do the work of five of the old ones by manipulating your tech path to increase the median value of available techs, and four can even do the work of six at some points of the tree.

They are not stronger. You are forced to research techs you don't need in order to increase the median. This is much more expensive than investing 25% beakers to block cheap techs. Beelines to a key tech are better than extra useless techs.

Of course deep beelines like Globalization and Rocketry and Hermitage were removed, so you can't even compare them anymore. Rifling is the only remaining traditional beeline, and the new RAs are clearly worse when you're trying for that.

 

[Furycrab]

The median is a type of average. In common English usage, we tend to assume that the unmodified "average" should be interpreted as "mean", but that isn't necessarily always the case.

Any overage from a Research Agreement is applied randomly to an available tech. How that overage is assigned is not yet clear to me. I have had it assign the overage to techs that were largely completed, so it's not locking out techs using the 25% + 3 mechanism it was using previously.

RAs are considerably stronger than they used to be if you take Rationalism and build the Porcelain Tower. You can easily make four of the new RAs do the work of five of the old ones by manipulating your tech path to increase the median value of available techs, and four can even do the work of six at some points of the tree.

Couple that with the increase in late-game tech prices, and the result is that Rationalism is the single most powerful policy in the game for peaceful play, and the Porcelain Tower is the strongest Wonder for that style. The Hanging Gardens is debatably stronger for a pure warmonger, since it can easily be translated to numerous additional early units that become fully promoted.

I kinda agree that with Porcelein tower and Rationalism RA are now a little stronger, but not by all that much.

Medians are like you said often considered the "mean average" but it's basically just taking a number.

So say you have a set of techs 1-3-3-6-18, your median is still just 3. (You just take the number in the middle, if you have an even number of Techs, you average the two in the middle) Later down the tech tree you have spots which open up quite quickly, but it will never "quite" go down as fast as with proper RA blocking was before since with RA blocking you effectively manipulated the value to always be the highest possible and block off entire sections of the tree. If you didn't RA block than ya, it's much faster.


I'm not a huge fan of Rationalism though. As a one point SP it's great, but with play testing with Piety in my last game (Persians), it's fairly crazy the difference in happiness/culture with Piety now <.< It will get the Science Victory faster, but at what cost <.<. I'm floating around with huge cities in a "wide" game with liberty.

 

[Martin Alvito]

They are not stronger. You are forced to research techs you don't need in order to increase the median. This is much more expensive than investing 25% beakers to block cheap techs. Beelines to a key tech are better than extra useless techs.

This is true for the early game. It is not true once you start talking about Renaissance techs. For Diplo/Science everything until the bottleneck (Scientific Theory, Biology, Steam Power) is more or less useless, with the possible exception of Economics for the extra GE slot and the Chemistry buff to Mines. Culture does care about Archaeology and Acoustics; the former must be bulbed if you want the GA slots early.

Another major advantage of the change is that you do not need to waste beakers blocking techs (Metallurgy) that you will never research.

The result is that you can clear techs for a Science win about as fast as you used to; the increased potency of RAs makes up for the extra techs that must be cleared. The Diplomatic win has been overly nerfed IMO; you need about as many techs for the win, but the saved via GE use on the UN doesn't make up for the necessary to win. It's hard to imagine a case on Deity where a Diplomatic win would ever be swifter or easier than a Science win.

Medians are like you said often considered the "mean average" but it's basically just taking a number.

Let me clarify. In mathematics there are many types of "average", the mean and median being two. In common English usage, the word "average" is interpreted as "mean", but in context it was pretty clear that the speaker meant "median" when saying "average".

 

[Furycrab]

Correct me if I'm wrong, don't you now need metallurgy for the Tech victory anyways? Unless you are comparing to the old way of getting the tech victory.

Strictly comparing though...

Old RA Gave you 100% of a random tech that isn't blocked. Assuming you did block you could have it always learning your most expensive tech, so you would get that many bechers every time.

New RAs, in the early game they are much weaker (no tower/rationalism) also there is no practical way of making it so that your most expensive tech is even close to your median.

Once you get past Renaissance, and get the Tower and Rationalism (for sake of comparison I assume you are getting it at the earliest convenience using GPs and what not)

The techs almost all the way to the tech victory come in sets of "levels" of 2 or 3 techs. Each new level is about 10-20% more expensive? (not at home was just eyeballing yesterday) The RA gives you 100% of median. If all your techs are on the same level you'll get the most expensive one, however you can't always make it so that your median is your most expensive tech, in fact most of the time your median will either be an average of two techs or your cheapest tech.

Even if we did have the old RAs I'm not convinced that we could use them the same way we did since they made many techs branch differently. You did have to spend points in techs you didn't need, however when done optimally, I'd still gather that you would get more bechers overall per RA than the new ones, but it was stupidly metagamey and many players hated it.

tl:dr version

New RAs are weaker than the old ones if you knew how to exploit the old ones, especially in the early game, however it will likely feel stronger/more practical/less gamey should you get the Tower / Rationalism.

 

[Martin Alvito]

Correct me if I'm wrong, don't you now need metallurgy for the Tech victory anyways? Unless you are comparing to the old way of getting the tech victory.

Yes, I'm comparing the former set of rules to the present ones here.

The techs almost all the way to the tech victory come in sets of "levels" of 2 or 3 techs. Each new level is about 10-20% more expensive? (not at home was just eyeballing yesterday) The RA gives you 100% of median. If all your techs are on the same level you'll get the most expensive one, however you can't always make it so that your median is your most expensive tech, in fact most of the time your median will either be an average of two techs or your cheapest tech.

The differentials start getting to 25% or worse once you get deep into the tree.

You're not leveraging the median properly. The "right" way to push your median is not to research purely horizontally. It's to have a laggard tech and a couple of advanced techs available when an RA concludes, use the RAs to hammer your way up through the techs above the laggard tech, then set up the scenario again using and the odd Great Scientist as needed. Juggling this is not easy, but all you really have to understand is how to set up tripping a series of techs at the points where the tree narrows.

You can't always pull that off; in those cases, you usually get four RAs to do the work of five of the old ones (by getting a combination of the value of the next level and the value of the mean of two levels) rather than getting RAs that tear through multiple low level techs and/or set up the next large RA series.

 

[Furycrab]

While I do agree that if we had the old RAs now, we probably couldn't get the same potential as we did before simply because it's harder to go about the new tech tree horizontally. The new ones are weaker in the sense that you don't get the number of bechers equivalent to your most expensive tech, in fact most of the time you'll get one of two scenarios 75% ish of your most expensive tech or 80-85%ish (Depending on if you have two techs open or 3). You also have to be fairly careful with Overflow because it can end up in Atomic Theory. Whereas the Old RAs you could not only get your most expensive tech, but you could also get subsequent ones leading to very ridiculous RA pushes.

I just don't get your "Getting 4 new RAs to do the same work as 5 old ones" when the old one let you do some fairly silly things. Many people have already pointed out that it's harder to manipulate a median, I can't see how you can make the new RAs produce more bechers than someone who properly managed the old ones.

edit: I do know how 4 will produce more bechers than someone who just let RAs go willynilly but you play on deity, so I'm assuming your comparing the pure potential of both.

 

[Martin Alvito]

It has to do with where the RAs trigger when.

As DaveMcV points out, the old RAs were unambiguously better in the early game. You could snag Education, Chivalry or Rifling with a minimum of bother (Steel was generally bulbed) by the 70's or early 80's in the case of Rifling. I agree completely here; my claim is that after the Renaissance and the Porcelain Tower, the RAs become better rather than worse.

Under the old system, you were forced to take laggard techs with RAs after the first wave of RAs. With the way the tree collapsed, you could not avoid taking Medieval techs with the second wave, or mid-Renaissance techs with the third. There simply weren't enough advanced techs available.

Under the median system where you have control of the (pre-overage) tech, you can leverage those chokepoints to compel the game to give you much higher per-RA counts after the first wave of RAs than you could under the old system. By the end of the first RA wave, you should have Education and Compass in hand, bulb Astronomy and finish the Porcelain Tower before the next wave begins to resolve. The net result is that the RAs after the first wave are far more effective, and this compensates greatly for the delay in reaching Education caused by the changes in the tech tree and the necessity of researching the bottom of the tree.

 

[Furycrab]

Well here's my issue, you mention the second wave of RAs around turn 110 being a bit weaker because you might have to put a medieval tech here or there, but even with meticulous planning it seems that no matter how you play your cards, the new RAs won't get you through it any faster.

What your telling me is not that the new RAs produce more bechers, much less 4 for the price of 5 old ones, but rather that the old system there was a cap in the Medieval/Renaissance where even if you got more RAs you might not be able to set it up where it's always your most expensive tech. I agree, but even if you did overspend and had an RA that had to be put into something "cheaper" because you couldn't quite put them in interesting techs, it will still be better then the second wave of RAs post patch where you might not even have unlocked Rationalism yet without a GS or two.

I'd argue that with the new RAs getting a science victory is a bit longer, but it's not the only factor.


How do you compel a chokepoint into giving you more bechers per RA than the old system. In the old system I would just block off everything cept the most expensive stuff and I'd make it so that each RA gets me the most expensive tech I can get. A choke is just as bad for the new system because your median will never be higher than the choke point. Heck it will likely be a lot lower until all your techs are in close relative value to the choke.

 

[Martin Alvito]

What your telling me is not that the new RAs produce more bechers, much less 4 for the price of 5 old ones, but rather that the old system there was a cap in the Medieval/Renaissance where even if you got more RAs you might not be able to set it up where it's always your most expensive tech. I agree, but even if you did overspend and had an RA that had to be put into something "cheaper" because you couldn't quite put them in interesting techs, it will still be better then the second wave of RAs post patch where you might not even have unlocked Rationalism yet without a GS or two.

You should be throwing all of your efforts at popping Rationalism and the Porcelain Tower before the second wave begins to resolve. I always RA Education and usually part of Compass even at the discounted RA rate, so that I can get the GS counter and Rationalism rolling faster. I also bulb Astronomy with either the Porcelain Tower GS or the 100 point GS.

How do you compel a chokepoint into giving you more bechers per RA than the old system. In the old system I would just block off everything cept the most expensive stuff and I'd make it so that each RA gets me the most expensive tech I can get. A choke is just as bad for the new system because your median will never be higher than the choke point. Heck it will likely be a lot lower until all your techs are in close relative value to the choke.

If you were truly pushing the envelope, you couldn't block fast enough and you lost a lot of turns to this problem unless you were fortunate.

What you want to do is find a chokepoint early in the tree and a branch point later in the tree, then hammer your way through the early chokepoint. For instance, Gunpowder is handy because researching it makes two techs of the next level available for research. Then you can hammer your way up to Economics through Machinery, Printing Press, Chivalry, Acoustics and Banking with just a few RAs if you play your cards right (assuming that you got to the Renaissance via Astronomy and have Navigation open).