Research tweaks?

[RocketSquid]

Many people are discontent with the current research system. This thread hopes to amass consensus and suggestions.

I don't really have anything right now, but other people hopefully will. In other words, this is a R&D discussion thread.

[Wjolcz]

Guess what I voted for?

I said it many times and I will say it again: the science-research system is all wrong. I don't need to visit the Mun to build myself a space shuttle.

Science should be actual science (also grant rep as a form of reward) and tech should be unlocked with money and time.

In this case tweaks are not enough.

[OHara]

I like the gradual unlocking of parts, because restricted choice forces creative solutions.  The fact that we need to do something (collect science) with our limited tech before getting the next level makes a fun challenge.

The need to remember to activate experiments is a bit un-fun, and there are suggestions to automate that. The silliness of carrying goo to biomes, and its disconnectedness to the technology that it unlocks, doesn't bother me too much, but it does lose my interest and makes it harder for me to remember the rules for collecting science.

It looks like experienced players can set up the game to unlock the tech-tree with money and time, using tree toppler and setting part-unlock costs in the new-game options, but I haven't tried that yet.

My best suggestion, which I've tried myself, to keep what we like about earning our way through the tech tree, is similar to an old suggestion 'actual R&D'.   I kept track of how many science points, reputation stars, and 1000s of funds, I gained using parts on nodes I had recently unlocked, and when this totaled the science points required for the next node, I considered the predecessor node to be sufficiently established technology, and opened the next tech-node by mod-F12.  (Book-keeping by hand was too tedious, and I was making up the rules as I went along, so I drifted to unlocking nodes when I felt I had gotten all I wanted out of the earlier tech.)

[Alshain]

I don't much care for the science, but the problem is people want to replace it with systems that would replace contracts too.  Contracts are fine, they have the benefit of not having to do stuff you don't like to do.  If you can find a way to replace the science without changing that then by all means go for it.

[Wjolcz]

8 hours ago, Alshain said:

I don't much care for the science, but the problem is people want to replace it with systems that would replace contracts too.  Contracts are fine, they have the benefit of not having to do stuff you don't like to do.  If you can find a way to replace the science without changing that then by all means go for it.

I disagree. They are not good at all. It would be better if they could be filtered like it is in Strategia mod. With each body having it's own tab.

EDIT: That would still be too random IMO, but still better than what we have now. I would like to see a proper step-by-step mission planing tool (so Mission Control can actually work as a proper Mission Control, not contracts slot machine). You pick a body and then each experiment/activity is selected also by you. Want to take only 3 tourists to the Mun and back and nowhere else? Want more? Less? It's up to you what you want to do in what order.

Seriously. Contracts are nowhere near good. Free will =/= randmoness.

Edited August 15, 2016 by Veeltch

[pandaman]

The problem with the current system is that you research what you need to get somewhere AFTER you have already got there.

Current KSP -

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Ok, I'll take a trip there tomorrow and see what I need" 

 (next day) - "Hmm,  it's a bit harder than I thought, I'd better design some landing legs and a ladder.  Oh! and bring a bit more fuel next time".

 

What it should be - 

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Oh s**t, how do we do that?  What do we need to find out?, What equipment and techniques do we need to develop to make it possible?"

 

A bit of an over simplification, but I hope you get the idea.

[Wjolcz]

36 minutes ago, pandaman said:

The problem with the current system is that you research what you need to get somewhere AFTER you have already got there.

Current KSP -

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Ok, I'll take a trip there tomorrow and see what I need" 

 (next day) - "Hmm,  it's a bit harder than I thought, I'd better design some landing legs and a ladder.  Oh! and bring a bit more fuel next time".

 

What it should be - 

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Oh s**t, how do we do that?  What do we need to find out?, What equipment and techniques do we need to develop to make it possible?"

 

A bit of an over simplification, but I hope you get the idea.

Exactly.

#fixnottweaks

[Alshain]

6 hours ago, Veeltch said:

I disagree. They are not good at all. It would be better if they could be filtered like it is in Strategia mod. With each body having it's own tab.

EDIT: That would still be too random IMO, but still better than what we have now. I would like to see a proper step-by-step mission planing tool (so Mission Control can actually work as a proper Mission Control, not contracts slot machine). You pick a body and then each experiment/activity is selected also by you. Want to take only 3 tourists to the Mun and back and nowhere else? Want more? Less? It's up to you what you want to do in what order.

Seriously. Contracts are nowhere near good. Free will =/= randmoness.

That's all fine as long as there is no penalty for doing what you want to do instead of what the game wants you to do.  The thread in your signature would suggest otherwise.  If I don't like to do tourism at all, then the game shouldn't force me.

If I'm going to lose reputation or money or whatever because I choose to play my game, then I'd rather just keep contracts and science as they are.

5 hours ago, pandaman said:

The problem with the current system is that you research what you need to get somewhere AFTER you have already got there.

Current KSP -

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Ok, I'll take a trip there tomorrow and see what I need" 

 (next day) - "Hmm,  it's a bit harder than I thought, I'd better design some landing legs and a ladder.  Oh! and bring a bit more fuel next time".

 

What it should be - 

Kennedy Kerman - "We are going to the Mun.."

KASA R&D  - "Oh s**t, how do we do that?  What do we need to find out?, What equipment and techniques do we need to develop to make it possible?"

 

A bit of an over simplification, but I hope you get the idea.

Yet in the real world it is a bit of both.  The Lunar Roving Vehicle only received funding days before the Apollo 11 landing.  Sure, there had been studies and other such things much earlier, but the project did not exist before we landed on the moon.

[tater]

"Contracts" as they are are absurd for a few reasons.

One, they are effectively nothing more than often bizarre, or frankly stupid side quests. I saw one last night to put a rover wheel on a suborbital trajectory. Seriously?

Two, why do third parties want "science" from the Mun, etc? I want to have one of the "contract" writing entities be your own scientists and engineers. Linus, Werner, etc. "We've gotten funds to do science on the Mun, which of these missions would you like us to pursue?"

Three, they have time limits that are meaningless given current game mechanics. The only time that passes in KSP is flight time.

Four, they are not player-driven.

 

I'm of the opinion that science, "contracts," and R&D all need to be tweaked in unison. 

I could see a research paradigm where you grab the node (different structure than we see now, for sure), and drag it to an R&D timeline. Upgrading the facility might allow multiple concurrent researches. 

The node would then have a sort of progress bar to competition over time. Many months. Oh, noes, players will warp. Good, that's the point, I want time to actually progress, anyway so that programs are not "What's a rocket?" to SSTO spaceplanes to Eeloo in 100 days. Having a node researching will then create a special class of engineering missions that pops up related to that node. Doing any of these reduces the R&D time. So playing the game, and building stuff can speed up those many month R&D times, or you can go about your usual missions until whatever is done. Funds might also be allowed to be expended to speed it somewhat. The R&D missions would be like parts testing, only better. Sure, the static test at KSC would knock a week off, but operating the test part in a far Kerbin orbit, then changing the plane of it might knock a month or two off it. We'd need sensible testing mission designs. Drills might require sample return to speed it up, crewed parts might require stations on orbit with crew for certain periods of time.

[Wjolcz]

1 hour ago, Alshain said:

That's all fine as long as there is no penalty for doing what you want to do instead of what the game wants you to do.  The thread in your signature would suggest otherwise.  If I don't like to do tourism at all, then the game shouldn't force me.

There are also a bunch of ways of how to deal with the "peanlties" you are talking about. The discussion and OP there is open to suggestions. I would love to hear yours too. The more people discussing it the better. And by the way, the thread there suggests and discusses something completely different than what you said about tourism. It's just that I'm not native english speaker and my wording might be poor.

28 minutes ago, tater said:

Two, why do third parties want "science" from the Mun, etc? I want to have one of the "contract" writing entities be your own scientists and engineers. Linus, Werner, etc. "We've gotten funds to do science on the Mun, which of these missions would you like us to pursue?"

Three, they have time limits that are meaningless given current game mechanics. The only time that passes in KSP is flight time.

Four, they are not player-driven.

I agree with this

28 minutes ago, tater said:

I could see a research paradigm where you grab the node (different structure than we see now, for sure), and drag it to an R&D timeline. Upgrading the facility might allow multiple concurrent researches. 

and this.

28 minutes ago, tater said:

The node would then have a sort of progress bar to competition over time. Many months. Oh, noes, players will warp. Good, that's the point, I want time to actually progress, anyway so that programs are not "What's a rocket?" to SSTO spaceplanes to Eeloo in 100 days. Having a node researching will then create a special class of engineering missions that pops up related to that node. Doing any of these reduces the R&D time. So playing the game, and building stuff can speed up those many month R&D times, or you can go about your usual missions until whatever is done. Funds might also be allowed to be expended to speed it somewhat. The R&D missions would be like parts testing, only better. Sure, the static test at KSC would knock a week off, but operating the test part in a far Kerbin orbit, then changing the plane of it might knock a month or two off it. We'd need sensible testing mission designs. Drills might require sample return to speed it up, crewed parts might require stations on orbit with crew for certain periods of time.

That's a great idea.

Edited August 15, 2016 by Veeltch

[RocketSquid]

I have an idea that I think will allow a more realistic research system without removing the motivation for return missions or causing too much work for the devs or new players.

• Science points (same as current) are produced at a constant rate by ground based labs at the R&D center, and at a slow rate by mobile processing labs.
• Raw data from experiments can be used to give a small boost to mobile labs, while funds can be used to do the same for ground based labs.
• Research gets faster as R&D gets upgraded, rather than the current system of limited-size nodes.
• New parts can be unlocked from a tech tree, which would basically be a re-organized version of the current one. This saves people from the inconvenience that may result if your needs change between starting research and finishing it.
• "Directed R&D" allows you to use funds in place of science for part of node costs. The cost per science grows the more science you try to replace, which means higher level nodes can get expensive fast.
• Using the Admin Building you can boost or throttle R&D in exchange for funds and possibly a little reputation.
• Unlocked parts go through an optional (via difficulty settings) testing phase. Using the part will speed this up, but during this time it has a chance of failing and a tiny chance of failing catastrophically.
• (The difference between failure and catastrophic failure is simple: A failing part ceases to work. A catastrophically failing part fails and damages other parts. As an example, an engine that fails will refuse to activate. An engine that fails catastrophically will explode.)
• Science experiments, part tests, and labs all have a chance of producing a breakthrough, giving a significant amount of science. The chances of a breakthrough for a given location and experiment go down for every time that experiment is run in that location. A general rule is that every new planet or moon will produce one breakthrough for the initial visits, and one per year if you build a proper scientific base.
• There would be new science labs, in 1.25, 2.5, and 3.75 sizes giving higher or lower research rates proportionally.

[Wjolcz]

12 hours ago, RocketSquid said:

I have an idea that I think will allow a more realistic research system without removing the motivation for return missions or causing too much work for the devs or new players.

• Science points (same as current) are produced at a constant rate by ground based labs at the R&D center, and at a slow rate by mobile processing labs.
• Raw data from experiments can be used to give a small boost to mobile labs, while funds can be used to do the same for ground based labs.
• Research gets faster as R&D gets upgraded, rather than the current system of limited-size nodes.
• New parts can be unlocked from a tech tree, which would basically be a re-organized version of the current one. This saves people from the inconvenience that may result if your needs change between starting research and finishing it.
• "Directed R&D" allows you to use funds in place of science for part of node costs. The cost per science grows the more science you try to replace, which means higher level nodes can get expensive fast.
• Using the Admin Building you can boost or throttle R&D in exchange for funds and possibly a little reputation.
• Unlocked parts go through an optional (via difficulty settings) testing phase. Using the part will speed this up, but during this time it has a chance of failing and a tiny chance of failing catastrophically.
• (The difference between failure and catastrophic failure is simple: A failing part ceases to work. A catastrophically failing part fails and damages other parts. As an example, an engine that fails will refuse to activate. An engine that fails catastrophically will explode.)
• Science experiments, part tests, and labs all have a chance of producing a breakthrough, giving a significant amount of science. The chances of a breakthrough for a given location and experiment go down for every time that experiment is run in that location. A general rule is that every new planet or moon will produce one breakthrough for the initial visits, and one per year if you build a proper scientific base.
• There would be new science labs, in 1.25, 2.5, and 3.75 sizes giving higher or lower research rates proportionally.

I like how you tried to create a compromise by putting in most of the needs and requests people have been asking for, but that still wouldn't cut it IMO.

Let me do the same:

• The tech tree is unlocked with time and money.
• Science is actual science that also grants you reputation for gathering it.
• The amount of reputation influences the monthly budget.
• The Administration Building gets proper programs that let the player choose the planetary body of interest (kind of like in Strategia mod). IMO the programs should progress in more or less this pattern: Scientific Programs (mostly unmanned probes, rovers, satellites, LKO/sub-orbital tourism, etc.) -> Colonization Programs (bases, stations, interplanetary vessels, tourism, etc.) -> Commercial Exploatation of Space (hauling ore from different places back to Kerbin to sell it, advanced interplanetary tourism, etc.). If a program is selected/active the tab relating to that program shows up in the Mission Control building.
• The Mission Control gets and actual Mission Planning Tool. Each planet/moon has it's own tab. When clicked on, the tab opens and all the possible scientific/commercial/tourism-related objectives roll out. Once that's out the player should be able to mark the ones he wants to complete around the selected body and accept the mission plan.

[tater]

Tech unlocked by time and money makes sense, and it also makes sense that particular tasks ("contracts" "missions," whatever) might speed things up. Certain science might also either speed things up, or actually be a requirement for tech to get past some milestone.

So imagine my idea above. You start research, which incurs some cost in funds, and might also have required precursors (in facilities, other tech, or even specific science). That node appears on a timeline with milestones every "Minmonth" (I use Minmus months because the munar month in KSP is 6 days, and a Minmus month ("Minmonth") is ~50). You get there by simple time progression, but perhaps each milestone could have a way to jump to it via completing science/testing. Note that some milestones might require the previous milestone, but others might not, letting you leapfrog.

Want to develop the Hitchhiker?  Say that one gets 4 minmonths of dev time. Since as a station/base part, rendezvous and docking will almost certainly be a thing, perhaps the first month can be jumped by completing a docking of 2 vessels. The third milestone might allow an orbital mission of at least a month, but no precursor, so you could launch that mission, and a month later you'll leapfrog and save 2 months. 

You get the idea.

[capitaniojr]

On 15/08/2016 at 2:59 PM, tater said:

"Contracts" as they are are absurd for a few reasons.

One, they are effectively nothing more than often bizarre, or frankly stupid side quests. I saw one last night to put a rover wheel on a suborbital trajectory. Seriously?

Two, why do third parties want "science" from the Mun, etc? I want to have one of the "contract" writing entities be your own scientists and engineers. Linus, Werner, etc. "We've gotten funds to do science on the Mun, which of these missions would you like us to pursue?"

Three, they have time limits that are meaningless given current game mechanics. The only time that passes in KSP is flight time.

Four, they are not player-driven.

 

I'm of the opinion that science, "contracts," and R&D all need to be tweaked in unison. 

I could see a research paradigm where you grab the node (different structure than we see now, for sure), and drag it to an R&D timeline. Upgrading the facility might allow multiple concurrent researches. 

The node would then have a sort of progress bar to competition over time. Many months. Oh, noes, players will warp. Good, that's the point, I want time to actually progress, anyway so that programs are not "What's a rocket?" to SSTO spaceplanes to Eeloo in 100 days. Having a node researching will then create a special class of engineering missions that pops up related to that node. Doing any of these reduces the R&D time. So playing the game, and building stuff can speed up those many month R&D times, or you can go about your usual missions until whatever is done. Funds might also be allowed to be expended to speed it somewhat. The R&D missions would be like parts testing, only better. Sure, the static test at KSC would knock a week off, but operating the test part in a far Kerbin orbit, then changing the plane of it might knock a month or two off it. We'd need sensible testing mission designs. Drills might require sample return to speed it up, crewed parts might require stations on orbit with crew for certain periods of time.

like the way you think, the science should be related to things done. and science parts should gain science over time on each biome, up to a max.

[Wjolcz]

On 19/08/2016 at 9:18 PM, capitaniojr said:

like the way you think, the science should be related to things done. and science parts should gain science over time on each biome, up to a max.

That would make sense as long as science points don't unlock the parts. Waiting for science points to build up just to unlock another node would be annoying.

IMO they should gather science (in a form of graphs and photos) that grants reputation.

[ShotgunNinja]

I believe the unlocking tech tree is not great, but fine and in line with many other games. Tech trees all work in this way from long time, is a proven system. It could be spiced up somehow, maybe rearranged. More mods (or even stock) could perhaps use 'progressive improvements' derived from the tech tree more often, that also make the part itself more interesting.

What would really make the science system 'more juicy' in my opinion however is improving the data collection, as that is central to the whole game. In fact I believe data collection is the main rationale that the player perceive, and it drive him forward as 'collect data today so that tomorrow I can collect more, better data'. Sound silly on paper, but not inside the brain.

In particular, I am of the idea that just having science collection and transmission 'take time' will improve the existing system a lot. This is suggestion number (1).

Also,situations are weak right now. There is the problem that most type of data you collect just doesn't fit with the 'landed/flight/space_low/space_high' dichotomy at all. So the situations should be extended.

At this point, if you want to extend situations, what possibly could be a situations? For start, a situation differ from another by some kind of environment factor. Consider the current system is evaluating one single environment factor: altitude. That is quite frankly barebone, and there is not much more in stock.

So more environment factors need to be added, before more rich situations can be added. This is suggestion number (2).

Now, both (1) and (2) taken in isolation don't add much. Buth together, they lead to a much interesting gameplay.

I'll give a practical example: there is the environmental factor 'radiation', and a set of situations depending on it are defined as 'radiation_low, radiation_high, radiation_extreme'. Also, data collection take days, weeks, or even months. And you have an instrument associated with radiation that consider those situations. All of a sudden you have a Van Allen Probe mission that is not 'role-play' but actual game mechanics.

Edited August 29, 2016 by ShotgunNinja

[AlexisBV]

Regarding things taking time to research, there's a mod for that. It's called KCT. It also adds a timer for building stuff, so no more insta-rescue missions, you have to plan ahead.

I tried it out, and though the UI could use some polishing it's a good mod. However I uninstalled it - I figured it added just a bit too much of a burden for my personal taste and my limited availability.

As for failures, I also tried out KSP TestFlight, another good mod. I ended up uninstalling it as well for the same reasons (and what I think are ridiculously high failure rates in the default config)

I agree that altitude/landed/not-landed is quite simplistic in terms of situations, but on the other hand you might never get enough science to unlock stuff if it's scattered over 65 thousand situations...

[nascarlaser1]

my only issue with the tech tree is the fact we should have statyputnik and the mk1 pod at the same time, so we can build probes before manned. I just want 1 part moved down the list, without the other parts like electric and the such being moved. they are fine where they are, but stayputnik should be moved, and the other probe cores could be reorganized, but I don't care bought those.

[Mad_Scientist]

I had an idea a while back, which is kinda similar to this topic.

The way to progress through the technical tree should be by doing experiments with a parts that you already had. For instance, in order to get to the orbit we stack a lot of small fuel tanks. But once you got there the scientist should think - oh, how we can do it better?? Lets increase the volume of the tanks!

You destroyed your first probe on the moon surface - 'hmm, if we would have something with higher impact resistance.. like a few sticks radially attached that can be extended'. Boom, small landing gear is available for the research.

The probe is going to an orbit, doing one turn and electricity is depleted. Scientists are working on solar panels from now on.

But then, what do we do with a science points. As I see it, there are few ways. First, is a similar of what is now - you spend scientific points to speed up the research of some particular part. Second, you require to recover/transmit science from specific biome in order to complete mission. It is more realistic, I think. Cause we are sending all the probes to specific places and we get money to do so.

[Sharpy]

On 15.08.2016 at 2:34 AM, Veeltch said:

I said it many times and I will say it again: the science-research system is all wrong. I don't need to visit the Mun to build myself a space shuttle.

 

And I said it many times and I will say it again, you don't need to visit the Mun to build a space excrementstle.

Not only you have other game modes and mods. You have a strategies that earns science for cash and rep. You have the initial science slider, and science reward multiplier - and you do earn science from contracts too. You have the labs, which can produce indefinite amount of science in LKO.

[Marzl]

Maybe the experimental parts already in the game can be used to spice up the tech tree a bit.

E.g. most tech tree levels can have one or more parts in it that are always marked as experimental first.

When you research such a level you get the contracts for this experimental parts auto-assigned. You can only proceed to next tech tree level when all experimental part contracts of prereq. levels are done.

Thus the contracts would get more significance, and you would actually need to test drive such a part to get the final version for your game. As they are not too much parts in stock the contracts can be hand-made so they are realistic. Maybe also connected to a difficulty setting (no/easy/medium/hard experimental contracts).

Edited August 30, 2016 by Marzl

[Sharpy]

 

Legit KSP game, no cheats, no mods (I couldn't be assed to remove them, but no mod features were used.)

Look at the clock - 24 minutes since start of the career. No object was launched, and none moved more than 2km away from KSC (one short excursion to reach Grasslands).

No minor biomes were harvested - no sub-biomes of Research Center, no Flag, not the Administration Center. Not the Shores either. The research tree can be completed within another half an hour.

This is my answer to "You can't progress if you don't go to the Mun".

[Wjolcz]

I honestly don't care anymore. The basic mechanics of this game will never change. Only more things will get added. Same thing happened to Minecraft.

[Pthigrivi]

I will say, complaining that science points ignore how research works in real life is a bit like complaining that health points in first person shooters ignore the medical realities of gun shot trauma. Continuing to run around after being shot multiple times certainly doesn't make sense, but immobilizing players and making it so they can't see or shoot straight until they convulse and pass out doesn't make for the best gameplay. Nor does micromanaging salaries and departmental funding levels amid ever-fluctuating budgets. Games require abstraction and simplification to be playable, and the need to produce constructive incentives, costs, and tradeoffs very often trumps strict realism. Having science points as a separate currency that can only be gathered by conducting science means you can't just satisfy other non-science-based game conditions and ignore experiments altogether if you want to expand the tech tree. This is a subtle but important mechanism thats actually really difficult to replicate using only funds. There are very legitimate concerns about the pacing and mechanics of gathering science, but fretting over how moon rocks make for better engines isn't a very compelling one. I'm much more interested in identifying what's actually going wrong and solving those issues, and with finding ways to introduce time as a factor without grossly destabilizing the game. Life support is probably part of the solution, but is a ways off and doesn't address time-warp abuse in the early game before a player has stations and multiple long-term concurrent missions. It probably mostly involves introducing more tightly calibrated deadlines and bonuses, and the mission planning tools necessary to meet them.

 

As an answer to the poll I would say science certainly needs work, and being one of the oldest and most critical career elements it should definitely be a priority. I simply believe there are accute solutions to what ails it that don't involve restarting from scratch. 

Edited August 31, 2016 by Pthigrivi

[Wjolcz]

Well, I really stopped caring about all this career nonsense, but here's a thing: KSP's science system doesn't allow for roleplay. And as you may, or may not be aware of sandbox games is all about roleplay. I guess you can just yell at me "THEN PLAY SANDBOX!", but the problem is this whole game is one big sandbox.

So let's look at this scenario: you start a fresh career save and what you want to do is a bit of roleplay. You have three options: go jets first, go unmanned into space first, or go manned into space first.

But wait... That doesn't seem right, does it? The real options are these.

1. Go manned into space first.

2. Go manned into space first.

3. Go manned into space first.

Want to go jets? Better go grab all that orbital science! Want to go unmanned first? Better get science!

And I guess you could just set the science reward sliders to the max, or Alt+F12 the whole thing (both options are no different). But what's the point of progression if you can do that.

And that's not the problem with the tree layout only. You would have to go orbital sooner or later to unlock more of it.

What I would like to see is a progression that when you decide not to leave the atmosphere and/or LKO you don't get so much funding to unlock the tree as quickly as you would by exploring other planets, but enough to keep your program running and doing what you've been doing so far.

But that will never happend.

So yeah, there's the sandbox for me that I've been playing since KSP 0.11. Or mods.

Edited August 31, 2016 by Veeltch