Killing off AI civs will affect your tech discovery rate

Tom DeMille · June 14, 2000, 11:08

While replaying OCC11, I made an interesting discovery.

After building United Nations (I had had no contact with the other civs until this point), I gifted all my tech to the AI civs. I was expecting my advance rate for new techs to speed up, but it didn't even budge!

What was going on? I thought the # of beakers required for your next advance is partially based on how far ahead you are of the other civs. This indicated otherwise.

The answer came later in the game. Two civs had been destroyed early on in the game, so there were 2 free AI slots. The remaining AIs started fighting during the space race, and Persia's capital fell. Civil war errupted and a new French civ was formed. Shortly thereafter the new Persian capital also fell, and the Japanese civ was formed.

With 6 AI civs, my rate of tech advance fell from an advance every 6 turns to an advance every 3 turns.

The tech discovery rate formula must take into account dead civs. Dead civs will of course be behind you science wise. So if you are going for an early landing on AC, don't let any of your opponents die.

Ming · June 14, 2000, 11:21

I still don't buy in into the fact that the other AI's effect your science rate.

I do understand that your science rate effects other civs, but not the other way around.

When I take a look at my science advance rate... it follows the number of beakers that have been discussed in past threads. And the number of AI civs, or their current level has never effected the number of beakers I need.

Can anybody really prove that.

Smash · June 14, 2000, 11:52

ai tech does,without a doubt,effect YOUR research rate.Look at the OCC logs.

I also got some unexplained drops in Fortnight 11.Had to be caused by something.

Aurelius · June 14, 2000, 13:28

I played OCC11 twice using separate strategies. In game one I had a 1960's result due to 4/5/6 turn science even though I had a fully developed science city under democracy. In game two I had a 1930's result due to a lot of 2 and 3 turn science even to the end.

I was very suspicious. But Tom's theory makes sense. In game one, two AI's were destroyed. leaving 4. In game two, one AI was destroyed leaving 5. And there was certainly no parity in science in either game--the one city ruled. So Something very strange is happening and I think there are more than one factor.

Though I know that science giveaway definately helps reduce the advance time or beaker requirements, I also noticed awhile back that one needs to give away science to most all the AI to get the benefits. So maybe the cut off is science parity with 5 of 6 AI and definately don't kill off more than one clan!

Nice theory Tom, I was so frustrated in game1 of OCC11!

SCG · June 14, 2000, 13:40

perhaps the game is trying to: 1)simulate how exhanging of ideas (trading techs) with other people can lead to new insights to old problems, leading to better understanding and more efficient research. or 2) competition between equals each trying to one-up the other, taking on the other's challenge, leading to more breakthroughs.

likewise, if you are far ahead, you may be more lackadazical in your research, knowing there is no challenge to keep up, and if you are far behind, you already see ideas in action and it is easier figure out how they actually work.

makes sense to me anyway

Tom DeMille · June 14, 2000, 14:09

quote:

Originally posted by Ming on 06-14-2000 11:21 AM
I still don't buy in into the fact that the other AI's effect your science rate.

I do understand that your science rate effects other civs, but not the other way around.

When I take a look at my science advance rate... it follows the number of beakers that have been discussed in past threads. And the number of AI civs, or their current level has never effected the number of beakers I need.

Can anybody really prove that.

Well, I had an autosave 2 turns before the "event" in my OCC replay, so did a little beaker counting.

Researching superconductor, I needed 3504 beakers.

Next turn, with a 7th civ having been created (with a full set of science), the beaker requirement had fallen to 1752.

I then gave away all my remaining science (laser, space flight, plastics) to all the AI civs, and the beaker requirements dropped again to 1679.

So something is going on...

Ming · June 14, 2000, 14:14

Sounds like pretty good proof to me

That's all I asked...

And that's why I love these forums, you learn something new every day!

shamrock · June 14, 2000, 17:51

Your right, Ming, these forums are great! Nice insights, Tom - thanks, I will definitely use this knowledge to help me beat Paul's record (he still holds the record, right?).

Ming · June 14, 2000, 18:07

If he doesn't hold the record, I will have a heart attack... GO PAUL GO!

Paul · June 14, 2000, 18:11

Yes, I still hold the record. I landed in 1645 in the second fortnight game (the one with the monarchy start). This is the only one of the fortnight games that was started on a random unmodified map, so all the results from the other fortnight games don't officially count as records.

Ming · June 14, 2000, 18:16

Paul... I'm sure you can beat any of the "other" records if you just sat down and tried it... I bow to the OCC master ...
(and I used to think I was good at it)

Paul · June 15, 2000, 00:42

Your research rate is affected by how far you are ahead of the AI. When I give all my technologies to the AI in my OCC games it certainly affects the number of beakers required for my next advance. When research takes three turns I can often get it back to two turns by giving away technology to the AI.

Hooha · June 15, 2000, 00:51

This is my first post in this forum; I've been lurking for several weeks, and learning a lot.

I have to agree that AI science has an effect on your science rate. I've seen it happen where I've given the AI several techs, checking with my science advisor between gifts, and seen the time needed to complete my research go down from 12 to 11 to 10, etc. I see no other possible explanation.

Carolus Rex · June 15, 2000, 02:47

The topic of this thread is why Adam Smith suggested using restarting eliminated civs for the Space Oddity Challenge (they keep half the techs).

Of course, not killing them in the first place is better, but how to control the violent AI?

Carolus

[This message has been edited by Carolus Rex (edited June 15, 2000).]

KhanMan · June 15, 2000, 06:29

This is wierd: back in the days when Strategy Guides were popular, I bought one for Civ2 and it gave a formula of how many beakers each advance cost. There wasn't any mention of ai civ's techs or lack there-of having any effect.

So I loaded up a saved game, and tried it. I build MPE, and gave every tech I had to all the other ais. My tech rate did not change before and after I gave these techs away.

Then I tried it in an OCC game (built MPE first, instead of Colossus). Same routine: gave away all techs, no change in tech rate.

Unfortunately, the manual only says: "As time progresses, new advances require more funding to research."

Anyone have a way to nail down this?

-KhanMan

Platty · June 15, 2000, 08:08

I remember from a strategy book that only on the highest levels does the gap between the AI and you affect the number of beakers required to gain the next advance. On the lower levels it's a straight increase. If I had the book to hand I'd say what the increases were.

vik · June 15, 2000, 08:58

quote:

Originally posted by Paul on 06-14-2000 06:11 PM
Yes, I still hold the record. I landed in 1645 in the second fortnight game (the one with the monarchy start). This is the only one of the fortnight games that was started on a random unmodified map, so all the results from the other fortnight games don't officially count as records.

Paul,

I think it's also important to point out that OCC#2 was the last saved game we used, as opposed to making a scenario. I think scenarios are much easier to get a better landing date because of the way new techs are presented to you to research.

Tom DeMille · June 15, 2000, 10:29

I did a little testing. I started a small map, 7 civs, deity, and made sure all civs had zero techs (edited them with cheat menu). I then started giving myself and others civs techs to see how my advance rate was affected.

The following table lists the beaker requirements for my first 10 advances. It compares the requirement for when all AI civs have no techs, 1 tech, 2 techs, 3 techs, and 10 techs

. . . . . . . . . . . . . All 6 AIs have

beakers for | 0 techs . 1 tech . 3 techs . 10 techs

1st tech . | . 10 . . . . . 10 . . . . 10 . . . . 10
2nd tech . | . 18 . . . . . 18 . . . . 18 . . . . 18
3rd tech . | . 27 . . . . . 24 . . . . 24 . . . . 24
4th tech . | . 36 . . . . . 36 . . . . 36 . . . . 36
5th tech . | . 45 . . . . . 45 . . . . 45 . . . . 45
6th tech . | . 66 . . . . . 60 . . . . 60 . . . . 54
7th tech . | . 84 . . . . . 84 . . . . 70 . . . . 63
8th tech . | . 96 . . . . . 96 . . . . 88 . . . . 80
9th tech . | . 108 . . . . 108. . . . 108. . . . 90
10th tech. | . 130 . . . . 130. . . . 120. . . . 110

Clearly the AIs have an effect on your beaker requirements, but the formula that is used escapes me

[This message has been edited by Tom DeMille (edited June 15, 2000).]

Ming · June 15, 2000, 10:39

Thank you... I was thinking of running similar tests to see what happens. Since I've been mistaken on this point for so long, I wanted to see what was really happening.

Good Job!

Tom DeMille · June 15, 2000, 10:47

OK, a little more testing, and even more bizarre behaviour. This time I gave 10 techs all at once to progressively more of the AIs.

. . . . . . . . . . . . . # AIs with 10 techs

beakers for | 1 AI . . . . .3 AI . . . 5 AI. . . 6 AI

1st tech . | . 10 . . . . . 10 . . . . 10 . . . . 10
2nd tech . | . 18 . . . . . 18 . . . . 18 . . . . 18
3rd tech . | . 27 . . . . . 27 . . . . 24 . . . . 24
4th tech . | . 36 . . . . . 36 . . . . 32 . . . . 36
5th tech . | . 45 . . . . . 45 . . . . 40 . . . . 45
6th tech . | . 66 . . . . . 66 . . . . 48 . . . . 54
7th tech . | . 84 . . . . . 84 . . . . 63 . . . . 63
8th tech . | . 96 . . . . . 96 . . . . 72 . . . . 80
9th tech . | . 108 . . . . 108. . . . 81 . . . . 90
10th tech. | . 130 . . . . 130. . . . 110. . . .110

You can see from the above table that by holding off giving anything to a single AI, my beaker requirements go down. This might explain some of the strange behaviour we see in OCC games. Sometimes beaker requirements will actually go up after we gift science to everyone.

The moral of the story... I'm still working on that.

[This message has been edited by Tom DeMille (edited June 15, 2000).]

Ming · June 15, 2000, 10:56

Gee... I wonder how all this applies to MP games. It could mean a BIG change in strategy if similar things are happening when AI's aren't involved...

Keep up the good work!

Aurelius · June 15, 2000, 17:47

quote:

Originally posted by Ming on 06-15-2000 10:56 AM
Gee... I wonder how all this applies to MP games. It could mean a BIG change in strategy if similar things are happening when AI's aren't involved...

You'd have to have science parity with 6 or 7 MP'ers to get the benefits-- according to Tom's chart and assuming it affects non-AI. I've heard MP'ers don't trade in general except with allies if the conditions are right.

Sten Sture · June 15, 2000, 18:53

But if you tech bomb an opponent with the GL, you could not only waylay their tech track, but reduce your own research time. This would probably work best in a 3 teams of 2 each type game where you could gift a unique tech outside your own team and allowing your own techs to stay separate.

This might be a terrible actual trade but...

Just for example: All six players the following five techs - Brz,Alpha,Laws,Burial,Monarchy

Team #1 has
A: Horse,Wheel
B: Writ,Lit and the GL

Team #2 has
C: Writ,Potry
D: Writ,Myst

Team #3 has
E: WarCode,Currcy
F: Writ,Mason

Teams #2 & #3 are concerned about player B getting Rep (for whatever reason) team #2 knows player B has been researching Rep for a couple of turns.

Player C trades Pottery to Player E vs Currency
Player F trades Masonry to Player D vs Myst

Player B now has Pottery, Currency, Masonry, and Myst for a total of eleven techs and instead of it taking him 72 beakers to get Republic it is going to take him 150 or so. Meanwhile Players C,D,E, & F get to research their ninth tech at a cost of possibly 90 vs their eight tech at a cost of 96.

Obviously, this is an overly simplistic example but what if someone with the GL is trying to get Monotheism or Democracy?

hehehehe.

Smash · June 16, 2000, 19:27

I am now fully convinced.I cheated up a "perfect" OCC spot with 2 whales and 2 wine and all rivers.I had no restart on.

2 civs were destroyed early and at NO point could I manage 2 turns per tech.Despite giving everything to ai and even at 1839 beakers I could not get 2 turns.Landing 1803

I replayed it with restarts enabled and got "normal" tech rates (2 turnsPer for most game) with significantly less beaker output.Landing 16something

I then played it again,only this time I left 2 civs uncontacted.I did get 2 turns but not for very long.Landing 1750 something.

So,keep ALL civs intact and make sure you contact all of them.The earlier the better.

I posted awile back wondering how this might effect MP but I don't think anyone beleived me.I have seen some "unexplained" changes in MP.

Tom DeMille · June 16, 2000, 20:16

Excellent set of tests, Smash.

I wonder if your tech rate is affected if you choose less than 7 civs at the outset of the game.

Scouse Gits · June 17, 2000, 02:42

I have a game going where the Barb activity is rampant. They come in at No. 5, in the top cities. So, I returned to a save when I know they destroyed the Babylonians. Yes, my advances went up a turn after this happened. Good spot Tom!

Smash - I think I saw a comment by you on another thread, when you mentioned that "test" results in Cheat Mode were different from open play. I am testing some aspects of the game - and I agree. Have you any more info?
----------
SG (2)

June 17, 2000, 04:19

I'm playing OCC#11 right now and noticed some strange swings in needed beakers which coincided with changes of my place on the powergraph from last to one but last and back. Coincidence?

kcbob · June 19, 2000, 17:46

Am I missing something?

If I give the AI some of my advancements so I can research my next advancement in a shorter amount of time, I've also just shortened the AI research time by my gift. Therefore, what good did it do me?

There ain't no way I'm give away Robotics to anyone for any reason.

What am I not seeing?

------------------
Frodo lives!

Better dead than "Red"... or green... or blue... or yellow... or orange... or purple... or white.

Aurelius · June 19, 2000, 18:25

Hi KCBob,

Haven't seen you post in a bit.

Yeah, giving away science like Robotics and Space flight doesn't make sense unless you're playing OCC. Most OCCers don't mind (in fact they like) the rate of science being increased for the AI since they expect to get a ship off to AC way before their AI rivals.

One example, Say you have 4 more years before you get your last sciences to finish off your ship--Fusion. And you have 3 more space parts to make before you can launch. It would be convenient to give away a bunch of science to the AI to facilitate a three year research of Fusion, thus timing the ship launch with the last year of research with nothing to spare.

But science give away is also good earlier in the game--as they say, 'before combustion.'

You'd think they have the best ribs on AC or somethin'

Sten Sture · June 19, 2000, 19:00

SS Arthur Bryant's prepared for AC launch, Sir!

kcbob - Lets say the AI is researching Invention and you are worried that they may build Leos before you. You know what techs they have from an embassy and see that they don't have Sea, Nav, Rep, BridgeBuilding & Theory of Gravity. You also think that they are landlocked and with Leo's you may be able to wipe out a bunch of their cities. So you call them up an give them all of the above techs - thereby increasing the number of turns before they get Invention. There is a thread in the archive called "A list of data" that runs down the research cost for each additional tech. The AI won't know that you are tricking him - and as a bonus - he likes you and trades maps with you, and you get your research time reduced on your current tech. Win, Win, Win.