Q about how game chooses Blind Research?

[Black_Phoenix]

I've notcied on several occasions that right in the beginning of a game I start researching tech that I didn't even specify in my priorities (e.g.: choosing build and discover and having the computer choose explore tech twice in a row.) I know that even if I focus in just one catagory I'll get tech choices from other catagories to make furthur research in the catagory of my choice possible. That I know. Here's my question: How does the game make it's chioce when it's time to research a new tech? I was taking a skim through the alphax.txt file and saw that each tech is given a numerical rating from 0 to 3 in each of the 4 catagories. Anyone know what equation the game uses? I'd check my strategy guide (I remember it having lots of equations in there that the game uses to determine stuff) but long story short...it was stolen I don't even remember seeing it in there anyway, so can anyone profer up some nuggest of wisdom? Thanks in advance!

Black Phoenix

[eisduotres]

You've just hit on a pet peeve of mine: I believe the "all knowing" AI (aka Artificial Idiot Savant) cheats on "blind" research on Transcend level, at least. I like to play Blind Research, but time and again I get the feeling that I'm being "steered away" from a beeline.

More than a feeling: I've tested thi in a pure smacx lineup playing the CC. Try to make a beeline to Brain Secrets, maxing energy, labs, building solars, etc., and 4 times out of 5 the Data Angels would beat me to it! And me with +2 research (I'm aware the AI gets quantitative cheats above Librarian, that's not the issue, which is tech tree steering)!

I've not seen any commentary on this here. Any thoughts?

[Ogie Oglethorpe]

I seem to remember hearing that a discovery in a given tech field makes it unlikely to have a second tech discovery in the same field. Say for example you are making a beeline through the builds to Industrial Auto. If you discover B1 Industrial base your next discovery will not likely be another build tech such as B1 social Psych or B2 Industrial Economics.

[David Byron]

My impression was that even with "Blind" research off you do not get to chose exactly which tech you want. I seem to remember noticing that there was often a choice missing. I was not able to see what the formula for working out what tech would be missing in a given situation was. This was with SMAC not SMACX.

[Chowlett]

The formula for working out what's missing is as follows:

MISSING_TECH=COMPLETELY_RANDOM

[Lord Maxwell]

Chowlett: Far from it. Most of the randomness in Sid games is actually most often simple "paper beats rocks beats scissors" series than can be reverse engineered. This could be the exception, but I pretty much doubt it.

The way in which techs for the non-blind choices is picked is atleast not random. (You get the same techs at every step, if you are anal enough you can tech in exactly the same order in two subsequent games. Given no tech from pods offcourse. Even though the manual says that you get a random subset of the set of available techs.)

[Jasonian]

Actually, Chowlett is right! If you have a certain number of techs you could choose, which ones you get is completly random!

[Lord Maxwell]

This is for non-blind research "randomness". Simple experiment verifies that it is lawbound, not random. I tried this with UoP, and I will give you first a list of all available techs, then the techs I could choose between. You will see that it is exactly the same in each of the four starts I did. (For the record, all starts were on normal map of planet, I always added the first pod to the first city and put the scout on hold.)

For the free tech the available techs are all first level techs except Information Networks, which is already owned.
All these techs were availble for choice.
I took Biogenetics every time.

Then I got a choice for next tech. The set of available techs are all level 1 techs except biogenetics and information networks, and Planetary Networks from the level two techs.
Of these all but Planetary Networks were available to choose from.
I took social psych.

For the choice of next tech the set of available techs is as above, plus secret and ethical calculus.
Of these all but industrial base were available. I always took secrets.

For the secret free tech the set is as above. The actual set to choose from is doctrine mobility, ethical calculus, planetary networks and industrial base. (Here we see alot of missing choices.)
I always took industrial base.

Now we get to where the set of available techs is: Applied Physics, Polymorphic software, Industrial economics, ethical calculus, doc: mobility, centauri ecology and progenitor psych.
Of these all but ethical calculus were available for selection.
I took industrial economics.

For the last choice the set is as above plus ind:automation.
The actual set is doc:mobility, centauri ecology, ethical calculus, progenitor psych, planetary networks and applied physics.
now again they are all available.

I invite ye faint if heart to play around with it. I will later conduct some research in the selection of blind research. It is simple to see that (atleast to a major part) the selection of the subset for selection out of the set of available techs for non-blind research is lawbound. There might be a random element in it, but I could not get it to materialize in the first few tech selections, so please try it yourselves.

[Chowlett]

Ok, in the light of that I will climb down slightly - but only slightly. You say that the techs missing are the always the same if you choose techs in the same order. But no-one ever will. And if there's no obvious formula to what's missing, just some arbitrary law (now there's an oxymoron) then you can't work out what will be missing unless you've been there before. In that case, the choice is effectively random.

[David Byron]

Oh absolutely not! Not random, just unexplored. I found it to be the same each time too, which suggests that if you want to make a direct path to some tech the order in which you persue the goal *might* be critical to how many pointless techs get in your way, and each pointless tech will slow you down a lot.

I wonder if the aliens get this same problem....

I wonder if with "random" tech you can get a tech which would be impossible under "choice"....

The generalised formula would be very useful. I would guess it is state dependent and how you get the techs doesn't matter, just what techs you have.

[zsozso]

I also strongly believe it is not random, but a rule-based or state-based algorithm, which is simply hidden from the user.

I played a large number of fast transcend games and have found, that to get Treshold of transcendence, the following techs are not prerequisite (plain SMAC):

D12 Singularity mechanics
B12 quantum mechanics
E13 graviton theory
E14 applied gravitonics
C15 controlled singularity

So, in theory, you could transcend without ever researching these 5 techs. However, in practice, you must research "E13. Graviton Theory" and "E14. Applied Gravitonics" to get the choice for "Treshold of Transcendence". Without these 2 techs, the threshold is missing from the choices. I have confirmed this rule in over 30 games! I used different research paths to got to the same point, but the result was the same, you must research those two Explore techs to get the choice for the Threshold.

At that point I developed a theory, that there is an extra (silent) rule, that you need to have lower level techs of the same type to get the choice for a tech and since Treshold of Transcendece is E15, you can't get it without having completed E13 and E14. Sounds logical doesn't it ?

I have never found this rule of thumb false in my games ever since in one sense: I never got a tech choice where I had prerequisites but missed all lower level techs of the same kind. On the other hand, I have found cases where I did have lower level techs of the same kind and the tech was still missing from the choices. So the rule maybe true, but it is not the only one, there must be some other(s) too.

Zsozso

[David Byron]

quote: Originally posted by zsozso on 02-07-2000 04:36 PM I never got a tech choice where I had prerequisites but missed all lower level techs of the same kind.

Did you mean to say "all" or "any" there?

Certainly in this frustrating random-stagnation game I'm getting eg. Cyberethics a level 4 build-tech before Industrial Automation, a level 3 build-tech (and also before Adaptive Economics, another level 3 build-tech). It's happening to the AI too, so there doesn't seem to be any tendency to pick up lower techs more often than higher ones with "random".

I think your observation about transcendence might be just a coincidence. There aren't many techs up that high so it could be just luck that the combinations open to you all banned Ascent, and as you say, the hypothesis requires something else -- there are some techs missing from the choice even among the first level techs for example. Also my impression of opening up old SMAC games under SMACX when I have eg a level one green tech missing (Progenitor Psych) is that the choice of high techs was NOT reduced because of this level-1 tech. I'm pretty sure I've chosen Eudaimonia while Progenitor Psych was sitting there, for example.

[BustaMike]

Zsozso, your observation may be correct for the Threshold of Transcendace but I am certain you do not need all lower level techs to research a higher level one. I have on many occasions discovered Environmental Economics (B5) before I discover Industrial Automation (B3). I have also discovered Air Power (E5) before discovering Doctrine: Initiative (E4). Maybe the game just requires a certain # of techs before it will let you transcend.

[zsozso]

Sorry, the use of all wasn't clear in the sentence quoted by David. What I meant is that if you are missing all B4 techs (there are 2 of them), then you wouldn't get the choice for any B5. So you do not have to get all B4s B3s and so on to get the choice for B5, but you need at least one of each previous level of the same kind.

But again I'm not sure about this rule for all cases, but there are some rules and it is not random I'm quite sure about that.


For the number of techs required to transcend: It is not a question of number, if you research any or all of the other 3 I listed above, they will not help, you still can't choose the Treshold. You specifically need to do the E13 and E14 techs.
[This message has been edited by zsozso (edited February 07, 2000).]

[MosesPresley]

I don't mean to steer this thread away from the main topic, but I thought this would be a good opportunity to criticize the tech tree structure a tiny bit.

Does anyone regularly discover Quantum Mechanics prior to Singualarity Mechanics? In all of my games over the past year, I have discovered singularity mechanics before quantum mechanics. To me this seems odd. Even if I play the game with blind research off, I find it difficult to discover quantum first.

Lastly since SMACX the tech tree seems a little more garbled than usual. Any comments?