Author Topic: RND in research time  (Read 17485 times)

Offline Eddie

  • Commander
  • *****
  • Posts: 560
    • View Profile
RND in research time
« on: November 03, 2016, 05:37:32 pm »
Is the randomness in the research time adding any value to the game? Because so far I can't think of any. And if it adds no value, why keep it?

I would prefer research to be more like manufacturing, where it displays how many scientist days it needs until completion so you can allocate accordingly. Right now I'm looking at my save files to see how many scientist days I need (It gives me more peace of mind, because I WANT TO KNOW!!!). Displaying that outright in game could be exploited because of the RND applied to research cost. You could simply cancel and restart a project to roll the RND again and get a better result. If the RND is gone, no more exploit and no downside to displaying that value.

Changing this would probably mean lower research times because of less wasted research days, but that can be adjusted in research cost.

Your opinion?

Offline Dioxine

  • Commander
  • *****
  • Posts: 5412
  • punk not dead
    • View Profile
    • Nocturnal Productions
Re: RND in research time
« Reply #1 on: November 03, 2016, 05:53:07 pm »
No. If anything should be random, it's research. I'd gladly increase randomness if research model was more intricate.

Cancelling would do you no good, since you have no way of telling what the random roll was. And you should NOT know, since research is never a linear and predictable process (unlike manufacturing). Also you'll lose the researched item if research requires it, and will have to spend another one to restart. You can naturally savescum to your heart's content, that's your power as a player, but no part of mechanics.

If you're obsessed with losing research days, research everything with 1 scientist. That way you get maximum time efficiency.
« Last Edit: November 03, 2016, 05:54:42 pm by Dioxine »

Offline Solarius Scorch

  • Global Moderator
  • Commander
  • *****
  • Posts: 11408
  • WE MUST DISSENT
    • View Profile
    • Nocturmal Productions modding studio website
Re: RND in research time
« Reply #2 on: November 03, 2016, 05:57:35 pm »
It's easy to say how much time will be required to produce a rifle, given the amount and quality of personnel and materials. Any factory which produces anything knows exactly how long it takes to complete the production process.
Now tell me exactly how much time would it take to research a rifle (assuming you don't know it yet), given the amount and quality of personnel and materials. :P

"Gentlemen, our country needs a fusion reactor. How long would it take to invent it?"
"Sir, with our current resources, that would be twelve years and three months."
"Excellent, we'll reserve the money needed to build a couple fusion reactors in the budget for 2029."

Sounds rather silly, no?


EDIT: I've been ninja'd, of course. :P

If you're obsessed with losing research days, research everything with 1 scientist. That way you get maximum time efficiency.

This is actually a very good point.

Offline Eddie

  • Commander
  • *****
  • Posts: 560
    • View Profile
Re: RND in research time
« Reply #3 on: November 03, 2016, 06:30:30 pm »
Also you'll lose the researched item if research requires it, and will have to spend another one to restart.

That part I've checked. No, you get the item back if you cancel a research project.

Offline Arthanor

  • Commander
  • *****
  • Posts: 2488
  • XCom Armoury Quartermaster
    • View Profile
Re: RND in research time
« Reply #4 on: November 03, 2016, 06:39:14 pm »
Indeed, discoveries are the least foreseeable thing since you never know when someone will go "Ah ah!" and get the breakthrough that will enable the project to be completed. Some things, like designing a new craft, could be planned, given existing knowledge of all the modules, since it is more engineering than actual research (in fact, that's what defense departments do: Open up a call for new something proposals when they identify a need,  with a deadline for the weapons companies to reply), but all the fundamentals like researching new technologies to try to understand how the hell they function should be highly random.

What would be cool is a system where you can assign interrogation of certain VIPs to help with a research project. Stuck on your flying development? Ask a guild engineer! Can't figure out voodoo? Ask an Esper!

Having some kind of not-normal, not-flat distribution that has a fairly reliable mean but also the possibility of outright getting stuck, coupled with a mechanic for getting unstuck (researching related tech, interrogating VIPs, reading books/data disks) would be a much better system.

But totally predictable research times, especially as we already have totally predictable breakthroughs ("research X and Y and that'll cover the pre-reqs for Z, now you can research it") makes no sense. Research sometimes fails or takes forever, and what if none of your brainer thinks of Z to make the research proposal, uh? Research times should be random and research completion should randomly open access to other research projects instead of 100% of the time.

Something like:

1 - Start research project X. It takes a random number of scientist-days, with a small chance every day that further progress is impossible. If you get stuck (and the lab should report it in the progress, like it does for Excellent and other statuses), go to * without cancelling the stuck project.
2 - Complete research project X. This was your last dependency for both project Y and Z.
3 - You get a chance of getting Y or Z as available projects, a chance of getting both and a chance of neither.
3a - You got Y, when finishing the research of Y, Z has a chance of becoming available. If you fail that, see *.
3b - Same but with Y and Z swapped
3c - You got both, lucky you!
3d - You got neither, see *.
* - Start another project. This can be VIPs, data disks, books, etc. that have no chance to fail and have a chance of "unsticking" research projects, and sparking new ideas (ie making a research project for which all dependencies are done but which wasn't available yet available). Different things can be set to only being able to spark and unblock certain projects (like an engineer won't spark or unblock voodoo, only construction related stuff, similarly, an esper won't spark or unblock construction, but will for voodoo. Old Earth Books can spark/unblock any early-mid game tech, Liber Occultus can spark/unblock voodoo or demon stuff, Data Disks can work for anything, etc.)

All the chances mentioned above would be influenced by the number of scientists working on the project: The more scientists, the more likely you are that one of them has an idea for a following project. The more scientists you have, the less likely you are of getting stuck. This becomes an incentive to put more scientists on a project, even though it will waste a few scientist-days on the last day.

Offline Dioxine

  • Commander
  • *****
  • Posts: 5412
  • punk not dead
    • View Profile
    • Nocturnal Productions
Re: RND in research time
« Reply #5 on: November 03, 2016, 07:03:09 pm »
This becomes an incentive to put more scientists on a project, even though it will waste a few scientist-days on the last day.

Excuse me, but you really need no extra incentives to throw more people at a project - it gets researched faster, and that's what counts. Truth be told, it's kinda crazy that you finish research faster for each scientist added, 1:1 - in reality, this isn't nearly as simple; to use a simple analogy, if digging a grave takes one man 100 minutes, throwing 100 people at the task won't result in grave being dug in a minute... However this all would require a complex research engine, and I'm not sure if XCOM is the right game for such an engine.

Offline Arthanor

  • Commander
  • *****
  • Posts: 2488
  • XCom Armoury Quartermaster
    • View Profile
Re: RND in research time
« Reply #6 on: November 03, 2016, 07:23:06 pm »
Yeah sure progress shouldn't be a 1:1 thing with numbers of researchers. And if it weren't, then the need for further incentives increases.

Also, faster completion is rarely an issue in Piratez or XCom Files, since you're more trudging along in the tech forest than eagerly looking forward to climbing the tech tree like in Vanilla (where getting, say, plasma weapons was a OMG I need this NOW, compared to researching stuff that unlocks stuff, or weapons that you don't intend to use because they, along with 5 other weapons, will lead to your next weapon). In both mods, researching 10 things at once and not wasting research days is often more valuable than researching 1 thing at a time 10 times faster and wasting research. Especially since research is sort of limited compared to vanilla where building another lab and hiring 50 more scientists is a pretty easy thing to do.

Using 100 people gets you 100 graves in 100 minutes, or 50 graves in 50 minutes if they can work in pairs. That kind of "parallel work" is pretty much what I did in Piratez and XCom Files. You only have few scientists, so losing even 1 scientist-day feels like a major loss. This is amplified by the "many short projects" situation in both mods compared to the "few long projects" of vanilla XCom, since you get more opportunity to waste brainpower.

This is especially true when tidying up lose ends and encountering the possibility of over-staffing since you have no idea how long a project is unless you look in the ruleset/your save. For example, using 20 brainers on projects that takes 5-15 brainers-day, twice, versus doing the 2 projects with 10 each: Waste 5-15 twice for a total of 10-30 and be done in two days, versus 0-10 total by doing both in parallel, with the possibility of getting one or both in a single day any ways.

So even in the current situation, there is a good reason to spread your resources on different projects.

Offline bladum

  • Colonel
  • ****
  • Posts: 213
  • Bladum
    • View Profile
Re: RND in research time
« Reply #7 on: November 03, 2016, 11:30:28 pm »
what about this

1) project takes 100 scientist days to be completed, which is unknown to the player
2) player assigns 10 scientist to project, which is known to the player
3) every day it adds 50% - 150% to the project (5 to 15 points), unknown to the player
4) "Progress" is being displayed
   4a) None <25%, no chance
   4b) Slow 25%-50%, no chance to have a daily breakthrough
   4c) Average 50%-75%, 10% to have a daily breakthrough
   4d) Good 75% - 100%, 30% to have a daily breakthrough
   4e) Excellent >100%, 60% to have daily breakthrough

Offline khade

  • Commander
  • *****
  • Posts: 509
    • View Profile
Re: RND in research time
« Reply #8 on: November 03, 2016, 11:57:05 pm »
By breakthrough, do you mean the 150% day or just finishing the thing?  Or something I didn't think of?

Offline Arthanor

  • Commander
  • *****
  • Posts: 2488
  • XCom Armoury Quartermaster
    • View Profile
Re: RND in research time
« Reply #9 on: November 04, 2016, 06:09:08 am »
Yeah, that would need to be clarified.

Also, the current random duration with fixed scientist  output is functionally the sea as fixed duration with random output. I'd agree that random output is maybe more like research, but if the effect is the same, it doesn't really need to be changed.

Offline Dioxine

  • Commander
  • *****
  • Posts: 5412
  • punk not dead
    • View Profile
    • Nocturnal Productions
Re: RND in research time
« Reply #10 on: November 04, 2016, 06:17:54 am »
I think Bladum had in mind something like in Master of Orion 2: as progress on an advance reaches a certain threshold, there is a daily chance of instantly finishing it. The chance steadily grows with % of nominal cost completed, but it takes a lot of research points for it to reach 100% (assuming that breakthrough didn't happen yet).

Offline bladum

  • Colonel
  • ****
  • Posts: 213
  • Bladum
    • View Profile
Re: RND in research time
« Reply #11 on: November 04, 2016, 09:27:24 am »
breakthrough is daily chance to have research completed and technology given to player.
which means with good luck you can have it research even at 50% but with bad luck even at 150% of progress.

Idea of this system is that you may have something low risk / high risk to research (optional new parameter for research). Most basic it may be used for low risk = human technology, high risk = alien technology.

So there are two random factors here:
  • how good scientist are performing - 50% - 150% of assigned people to project
  • how risky is the area of being researched, there may be different threshold levels for different research. high risk = more curvy, low risk = more linear. Above i just gave example.
« Last Edit: November 04, 2016, 09:30:07 am by bladum »

Offline yrizoud

  • Commander
  • *****
  • Posts: 1014
    • View Profile
Re: RND in research time
« Reply #12 on: November 04, 2016, 10:57:22 am »
The original game had a system for research efficiency (100 people for one day not as efficient as 1 person for 100 days), I guess OpenXcom already implements it the same way.

Offline Solarius Scorch

  • Global Moderator
  • Commander
  • *****
  • Posts: 11408
  • WE MUST DISSENT
    • View Profile
    • Nocturmal Productions modding studio website
Re: RND in research time
« Reply #13 on: November 04, 2016, 04:09:32 pm »
Bladum, I would actually like a more complicated system, but it would be a huge change. Still, let me expand on it for fun:

Science teams!

A science team is a group of researchers working on a particular project. It is actually a roster with several positions. In order for it to make sense, scientists would need special skills, like in X-Com Apocalypse: Science (research of their own field) and Administration.

A team consists of:
- Project Leader. The most important person. Their Science skill generates research points as normal. They also increase the whole team's research output by a certain percentage, depending on Administration. You can only have one in a project.
- Project Coordinator. Only generates half the research points they would as a leader or Researcher, but slightly increases Researchers' efficiency with their Administration skill. The more Researchers, the lesser the bonus. You can have as many Project Coordinators as you like.
- Researcher. Simply generates research points with their Science. Doesn't use Administration. You can have as many Researchers as you like.
- Lab Technician. Doesn't actually generate research points, but uses their Science skill to prevent research setbacks (the lower the skill, the greater the chance of something going wrong). Doesn't use Administration. You can have as many Lab Technicians as you like.

I suppose it's rather poorly designed at this point, math-wise, but it's just for fun, right?

Offline Arthanor

  • Commander
  • *****
  • Posts: 2488
  • XCom Armoury Quartermaster
    • View Profile
Re: RND in research time
« Reply #14 on: November 04, 2016, 06:20:04 pm »
Hehe but then you'd end up putting scientists with good administration skills as project leaders and that'd be unrealistic;P

Project leaders should boost other scientists by their science score and coordinators by their admin score. That'd be more realistic. Project leaders come up with grand ideas and are inspiring others, the coordinators actually are the pivotal support people. I like what you did with techs (or research assistants?). Having good people there does make the research much more dependable.