Rebalancing Proposal

[Thalassicus]

The AI spending more on tile purchasing should reduce those large AI gold surpluses we've been concerned about. The AI seems to handle tile purchasing rather intelligently, getting tiles in cities with A) large populations or B) near enemy borders.

We can rebalance Washington if necessary. He ranked #1 in the favorite leaders poll so I'm not too concerned. I've copied below the the costs for purchasing a hill or forest. As Seek pointed out, most people probably get production tiles like hills and forests, since cultural expansion reaches those more slowly.

 

[Txurce]

That looks good on both fronts. I wasn't sure the AI bought tiles - although it certainly seemed like it. Your confirming it eliminates the concern.

The current difference between America and a generic civ in tile costs is bigger than when I last played, and looks significant enough from a distance.

 

[Zaldron]

v131.12 again, this time my initial thoughts on the new RAs.

I was shocked to discover that an industrial age RA now costs 3000g. Just as an example, my current net science income is 500beakers/turn while my current net gold income is about 150/turn. Let's grant that the AI is cranking TWICE my beakers/turn (generous guess). In that case I'll get 3% of 1500 times 30 turns beakers back out. This works out to 1500 * 0.03 * 30 = 1350 beakers over 30 turns for 3000 gold up front. Given that I'm making three times as much net science compared to gold I'm pretty surprised that the return on an RA is so bad (although I like the implications that the AI can burn their gold on it).

 

[Thalassicus]

In a game several months ago I saw the AI buy literally a dozen tiles around a city on my borders. It was really nice because I had my army there and planning an invasion, so they were indirectly expanding my borders. Overall they do seem to make frequent purchases - and usually not so extreme.

Remember that Declarations of Friendship double RA income, in your case to 2700, close to the 3000 cost. If the cost increase was too much I can tone it down again... perhaps a 1500? If I remember right I basically set it to [20% * cost of techs in era].

I generally balance from gross income instead of net profit. Profit is much more volatile than income. Consider if we're making +1000/turn with -950/turn expenses. If we enter a golden age our income might go up to 1100. This is a mere 10% increase in income, but a 300% increase in profit.

 

[Txurce]

Remember that Declarations of Friendship double RA income, in your case to 2700, close to the 3000 cost. If the cost increase was too much I can tone it down again... perhaps a 1500? If I remember right I basically set it to [20% * cost of techs in era].

Why don't we get a little more feedback? Since you raised RA costs (my last game was official v131) I'm surprised by how many RA's I still sign... but most of those were with DoF's, and presumably with more gold available than there is now in the betas.

 

[orangecape]

I was thinking about a formula for figuring out how much an RA should cost by era. First off, I will assume that we have a DoF 2/3 of the time and the AI has the same science rate as us, so roughly 10% increase in science. 10% * 30 turns is 300% of a turn's worth of science. For a science win in a 280 turn game we probably complete 64 techs, maybe 8 of which come from free sources or GS bulbs, so 56 techs. Over 280 turns we are averaging 5 turns/tech. If these assumptions hold then a RA is worth 60% of a tech.

The math can be seen below for most reasonable values using the assumptions above:

Turns/tech(game length): beakers gained from RA

10 turns/tech(560): 30% of a tech
9 turns/tech(504): 33% of a tech
8 turns/tech(448): 37.5% of a tech
7 turns/tech(392): 42.8% of a tech
6 turns/tech(336): 50% of a tech
5 turns/tech(280): 60% of a tech
4 turns/tech(224): 75% of a tech
3 turns/tech(168): 100% of a tech

Assuming all of that is true then pricing an RA at 20% of a tech should be fine. In a 280 turn game we would expect to get 67% more beakers than we spent in gold.

In the example above the tech was costing 15k and the player had 500 science per turn so they were getting 30 turns/tech, which gives the player only 15% of a tech. The trick is that the RA is presumably priced at the median part of the era and the player may well be only one tech into the era. However, if people are signing RAs at random times through the era I am sure that 20% of median tech cost should be all good.

 

[Jaybe]

v131.12 again, this time my initial thoughts on the new RAs.

I was shocked to discover that an industrial age RA now costs 3000g. Just as an example, my current net science income is 500beakers/turn while my current net gold income is about 150/turn. Let's grant that the AI is cranking TWICE my beakers/turn (generous guess). In that case I'll get 3% of 1500 times 30 turns beakers back out. This works out to 1500 * 0.03 * 30 = 1350 beakers over 30 turns for 3000 gold up front. Given that I'm making three times as much net science compared to gold I'm pretty surprised that the return on an RA is so bad (although I like the implications that the AI can burn their gold on it).

Don't forget that you also get a bleep-load of beakers at the conclusion of the RA due to Thal's not being able to disable it.

Unless of course, he has!

 

[Zaldron]

Why don't we get a little more feedback? Since you raised RA costs (my last game was official v131) I'm surprised by how many RA's I still sign... but most of those were with DoF's, and presumably with more gold available than there is now in the betas.

That was actually part of my motivation for posting: To get more feedback, especially from wide empires with a lot more science to multiply and gold income to spend on RAs.

I know that net gold income is very volatile, but I feel it's more meaningful to use it in comparisons (possibly averaged over turns) because you're comparing against net science. If science had a "research progress upkeep" that reduced net beakers each turn (to maintain facilities etc) before getting applied to new research then we could more accurately compare gross incomes.

All that said I'm still not 100% sure it's overpriced. I just know that in my current game I'm around 30 turns from a culture win having signed exactly one RA, because I was buying buildings and CS alliance.

 

[Seek]

All that said I'm still not 100% sure it's overpriced. I just know that in my current game I'm around 30 turns from a culture win having signed exactly one RA, because I was buying buildings and CS alliance.

That's a good point - RAs may not be overpriced in the lategame in terms of beakers gained, but in comparison to other things one can spend one's gold on it may not be worth it.

Agree with your point about net vs gross income also, btw.

 

[Txurce]

That's a good point - RAs may not be overpriced in the lategame in terms of beakers gained, but in comparison to other things one can spend one's gold on it may not be worth it.

Agree with your point about net vs gross income also, btw.

I drew the same conclusions from what Zaldron wrote: that one way or another, RA's weren't the best use of his gold. Up to a point, that's okay; beyond that, we're losing choice, rather than creating it. I'm looking forward to playing next week.

 

[EmperorFool]

I think we need to strike a balance between automatic signing of RAs and them never being worth it. In my previous games I would sign every RA because they were so cheap and I had so much gold. In my current game I'm warmongering so I also sing every RA I can get--both of them--because no one is willing to sign one. I'm still on an early 131 beta (.3?) and haven't played much since the weekend.

The way Zaldron describes it, maybe we've swung too far in the opposite direction where they are so expensive you can only sign one at a time and/or signing them just isn't worth it. It's really hard to know since each new beta changes many things at once.

In my work building a website we have been doing A/B testing lately. That means that as users hit our site we show them one of two versions: A or B. If we change more than one thing in each version, there's no way to know if the increased traffic/page-views are the result of one or the other change. Thus it's important that we change just one thing and one thing only between each version. Otherwise it involves too much guessing.

To translate that to Civ5 and VEM, perhaps we need to let things simmer down between changes to allow for more experience to help decide whether change X or change Y is the reason the game suddenly became harder/easier.

 

[Txurce]

To translate that to Civ5 and VEM, perhaps we need to let things simmer down between changes to allow for more experience to help decide whether change X or change Y is the reason the game suddenly became harder/easier.

This has always been my philosophy. In the present case, however, I think we are better off completing the major planned changes first. That will give us a version that includes the two major changes - AI gold spending and Opportunities - along with the various nerfs. Otherwise, I think we'd duplicate a lot of testing.

Of course exceptions can be made for specific changes that are easily identifiable as not working for one reason or another.

 

[Thalassicus]

Each beta is generally a flurry of activity to implement ideas under discussion in the weeks (or in this case months) preceding the beta. After the beta's released, there's a few more weeks with a lower rate of changes to get a feel for how things play out and make smaller adjustments. This downtime also gives me an opportunity to focus on other things.

 

[wobuffet]

Tiles are quite a bit more expensive now, and it's definitely more challenging in a fun way.

If adjust the price changes, I'd say make the 2nd-ring tile a bit less expensive in terms of both and (the CS I just conquered on turn 194 only had 3 tiles filled out beyond the 1st ring!), but keeping the 3rd-ring tile prices about as high as they are now.

Do we have any idea how intelligently the AI (esp. America, I guess) buys tiles with gold?


Also, count me in support of the beta flurry, if it matters.

 

[Zaldron]

I just finished my first game with v131.12 so I'm going to just make a few observations. I had a blast with this version in spite of a few quirks I'll note below. I played a tall peaceful empire on standard/standard/emperor. My final cities were about 36, 31, and 30 pop. I ended with ~140 happiness.

I finished with a cultural victory at turn 293, which is still fairly early. What's truly impressive is that I only had broadcast towers up for THREE turns before I finished the sixth tree, and that only because I bought them outright with gold (no doubt in my mind that three broadcast towers > a single RA for the gold; more on that later).

Although both culture and science were slowed in this version, culture was slowed much less. I had one or two techs in the modern era before I won culturally, and the entire bottom (dynamite) row in industrial was still *not* filled out at all. I actually don't know if I would have been able to win a science victory before I ran out of time, even with the *strong* possibility I would fill out an entire 7th tree (enlightenment) for the bonus techs.

* At the beginning of the game I got two unit opportunities and then no more opportunities of any sort through the entire remainder of the game.

* Tile expansion at the beginning seemed *slightly* slow, see my previous post with those observations. I suspect that increasing the rate while slightly reducing the tradition policy would be perfect.

* RAs cost 3000g in the industrial era, which may or may not be balanced, I'm not sure yet. But in the modern era they cost 6000g which is definitely not balanced as my net income actually dropped from industrial to modern due to increased building and unit maintenance but no additional sources of income (you don't get any upgrades to villages and possibly one additional merchant). As a side note, I think domestic trade routes were only about 5% of my total gross income (which may be perfectly fine since it's so passive). If a better number is needed I can load a saved game and check.

* Techs were taking about 13 turns on average to research and a GS would knock off...two turns of that. I know that bulbing a whole tech is definitely OP but for two turns of research I'll pop it for a golden age (or settle it if it pops early enough). On a similar note I think tech costs may have increased too much. My pace seemed *much* slower than in previous versions and the AIs were running over me in tech. I'd love to hear anyone else's experience with this.

I may try to play a game where I shoot for more cities and try to work towards a science victory but I haven't figured out my next game yet.

 

[Txurce]

* RAs cost 3000g in the industrial era, which may or may not be balanced, I'm not sure yet. But in the modern era they cost 6000g which is definitely not balanced as my net income actually dropped from industrial to modern due to increased building and unit maintenance but no additional sources of income (you don't get any upgrades to villages and possibly one additional merchant).

* Techs were taking about 13 turns on average to research and a GS would knock off...two turns of that. I know that bulbing a whole tech is definitely OP but for two turns of research I'll pop it for a golden age (or settle it if it pops early enough).

In my opinion, these (6000g RA's, 2/13 GS) are two examples of adjustments that seem clearly off, regardless of the overall balance, and could be shifted somewhat before we're done with all the planned changes.

 

[Seek]

I just finished my first game with v131.12 so I'm going to just make a few observations. I had a blast with this version in spite of a few quirks I'll note below. I played a tall peaceful empire on standard/standard/emperor. My final cities were about 36, 31, and 30 pop. I ended with ~140 happiness.

I finished with a cultural victory at turn 293, which is still fairly early. What's truly impressive is that I only had broadcast towers up for THREE turns before I finished the sixth tree, and that only because I bought them outright with gold (no doubt in my mind that three broadcast towers > a single RA for the gold; more on that later).

Although both culture and science were slowed in this version, culture was slowed much less. I had one or two techs in the modern era before I won culturally, and the entire bottom (dynamite) row in industrial was still *not* filled out at all. I actually don't know if I would have been able to win a science victory before I ran out of time, even with the *strong* possibility I would fill out an entire 7th tree (enlightenment) for the bonus techs.

* At the beginning of the game I got two unit opportunities and then no more opportunities of any sort through the entire remainder of the game.

* Tile expansion at the beginning seemed *slightly* slow, see my previous post with those observations. I suspect that increasing the rate while slightly reducing the tradition policy would be perfect.

* RAs cost 3000g in the industrial era, which may or may not be balanced, I'm not sure yet. But in the modern era they cost 6000g which is definitely not balanced as my net income actually dropped from industrial to modern due to increased building and unit maintenance but no additional sources of income (you don't get any upgrades to villages and possibly one additional merchant). As a side note, I think domestic trade routes were only about 5% of my total gross income (which may be perfectly fine since it's so passive). If a better number is needed I can load a saved game and check.

* Techs were taking about 13 turns on average to research and a GS would knock off...two turns of that. I know that bulbing a whole tech is definitely OP but for two turns of research I'll pop it for a golden age (or settle it if it pops early enough). On a similar note I think tech costs may have increased too much. My pace seemed *much* slower than in previous versions and the AIs were running over me in tech. I'd love to hear anyone else's experience with this.

I may try to play a game where I shoot for more cities and try to work towards a science victory but I haven't figured out my next game yet.

I played 2/3 of an aggressive game on v131.12 today, and found culture border costs exorbitant (I wasn't ignoring culture either!) with some cities waiting upwards of 100 (!) turns after just a few tile expansions - my "perfect" city locations were less than ideal, to say the least. Buying tiles became a necessity but the cost, oh the cost.. I think somewhere around the halfway point between these and v131's levels would be good, it's just no fun entering the lategame with only a few tiles in the second ring!

As a side note, on the other continent the Inca and Babylonians did not seem to be affected by the higher culture costs - both went Tradition->Piety while I went Liberty/Honor - and had borders out to the fifth ring in all directions. These two civs seemed to be playing v131 in terms of tech also, as both were entering Industrial as I was entering Renaissance! (Admittedly, I did neglect picking up Education for far too long.)

I thought the speed at which RAs costs' rose was too sharp as well (1000/Ren, 3000/Ind, 6000/Mod; don't know the price of earlier ones as I didn't get any before Ren because I was too busy killing everyone on my continent) - since one should always aim for getting RAs with the tech leaders, it was downright prohibitive, going against the general idea that VEM's modified RAs are intended to equalize tech. The general tech rate was pretty spot-on, I thought (if only those runaway civs weren't so far ahead) - it's looking like the game should finish up around t350 or so - but wouldn't be opposed to scaling it back slightly (I did find myself with too few building options sometimes).

Great Scientists haven't been modified since I brought them into VEM over a year ago and tech costs have gone up at least three times since then, IIRC - I fully agree that they should be made stronger.

I also got the Scout opportunity and no others following.

 

[Thalassicus]

the CS I just conquered on turn 194 only had 3 tiles filled out beyond the 1st ring!

Citystates do not purchase tiles or build much in the way of cultural buildings. They do however have a "minor civ plot culture cost multiplier" variable we can modify.

I'm unable to access the game right now, so could you test out good values for the Great Scientist seek? If they seem to give a good amount of science in the early game, we probably only need to change the exponent. The numbers are located in:

Vem/Cities/VEC_SpecialistsandGP.xml

Here's some possible changes for v.15:

Cities

• Reduced the culture border expansion exponent to 1.3 (was 1.4).
• Reduced the cost increase per tile purchased to 8 (was 10).

Policies

• Halved the effect of Landed Elite's bonus to border expansion.

Research

• Halved the cost of research agreements.

 

[Seek]

Citystates do not purchase tiles or build much in the way of cultural buildings. They do however have a "minor civ plot culture cost multiplier" variable we can modify.

I'm unable to access the game right now, so could you test out good values for the Great Scientist seek? If they seem to give a good amount of science in the early game, we probably only need to change the exponent. The numbers are located in:

Vem/Cities/VEC_SpecialistsandGP.xml

Sure! I should have some time tomorrow.

Here's some possible changes for v.15:

Looks good!

 

[wobuffet]

I thought the speed at which RAs costs' rose was too sharp as well (1000/Ren, 3000/Ind, 6000/Mod; don't know the price of earlier ones as I didn't get any before Ren because I was too busy killing everyone on my continent) - since one should always aim for getting RAs with the tech leaders, it was downright prohibitive, going against the general idea that VEM's modified RAs are intended to equalize tech.

Well, rapid increases in the price of Research Agreements actually have an anti-runaway mechanic built in: if I'm an era or two ahead of everyone else, each of my RAs will cost a lot more than other Civs' will (esp since AI civs that are behind will demand Gold before signing one with me).

That said, yes, the late-game RA prices are probably a bit too high (a 50% reduction seems too big). Either way, they're definitely an improvement from pre-v131 prices though, in my opinion.

Citystates do not purchase tiles or build much in the way of cultural buildings. They do however have a "minor civ plot culture cost multiplier" variable we can modify.

Can we turn this multiplier down then? Or maybe my experience was a one-off... has anyone else noticed a mid-game CS with very little territory to its name?


Re: the culture discussion above, I agree that culture victories still seem too fast; in my current game, India's got 28 policies filled out by turn 220. Policy costs could stand to be scaled up a bit more, at least for the later (the 20th and on, say) ones.


Also, I'm very glad to report that I'm not finding myself awash in Gold by turn 220ish in my current game! The AIs are still managing to rake it in (917 net gpt for Arabia and 353 for India by turn 220, somehow... even with reduced maint, I don't know how that's happening...), although my finances would be in better shape had I kept a few more puppets for gold purposes – I razed a few medium-sized cities to keep my Happiness up (most of my Science production is in my Capital, so I've been trying to keep that bonus high).