More intutive science: Samples=return, data=transmit, one shot, no diminishing return

[allmhuran]

We've all seen the issues with 0.22 science, and most of us have seen that Squad is looking at changing things in 0.23. But from what I've seen the changes that have been proposed right now are still pretty unintuitive. Transmission logic is "being redone", mystery goo will behave differently and very strangely - I can't find the post about this.

I think I have a much simpler, more intuitive system that still addresses allof the concerns with 0.22 mechanics.

1) Every science generating activity grants 100% of its science value. There is no such thing as transmission loss.

2) No diminishing returns. No experiment can be repeated. If you attempt to perform the same experiment you've done before, the game should simply say "you already have this data" (or show the data but also show this message).

3) Samples contain two experiments: the sample itself, and data about the sample:

3a) Samples themselves must be returned, no transmission is possible for samples. This is a high value science objective.

3b) Sample data is granted on return OR by processing the sample at the new lab module. This is a medium value science objective.

4) Any non-sample experiment (crew reports, EVA reports, temperature, pressure, etc) can be transmitted or returned for 100% value. No transmission loss. This is a low value science objective.

5) Mystery goo counts as a sample that you take with you, not data (and so must be returned for the sample value, and/or processed at a lab for the sample data value). Each container can only be used once ever. If you already have that particular data you should be informed of this and not perform the experiment (so as to not waste the container).

To me, this seems to make a lot more sense.

Right now it's not at all intuitive that things which are pure data (like temperature readings) would have diminished value when transmitted, and it's also not clear why any data for samples is returned via transmission.

The fact that each experiment only grants science points once ever, with the complete removal of diminishing returns, removes any kind of tedious grinding from the game. Yes, I know repeating experiments is important in the real world.

High value science objectives are the most difficult (requires sample return).

Medium value science is somewhat difficult (requires that you return or carry a lab with you).

Low value science is easy (only requires having enough energy available to transmit).

The only hesitation I have with this system is that only being able to perform any experiment one time (or at least, only getting science points one time) might be seen as making career mode "end" too soon. But I don't think this is actually the case. With the addition of more biomes there should still be easily enough experiments available to keep people busy without forcing them to redo previous experiments. I'm sure more experiments will also be coming along, like atmospheric samples, more data experiments, etc. And I'm sure there's much more to be added to the overall career mechanics outside of science collection as well.

Edited November 19, 2013 by allmhuran

[Kasuha]

I have already proposed similar thing in the past, with the following differences:

- each experiment has "transmission" and "delivery" science value. If you just deliver, you get both. If you transmit, you get just the transmission value. If you then deliver, you get the "delivery" part.

- repeating a measurement is always possible but you only get science point for what was not yet transmitted/delivered

- manned/unmanned counts. Probes get less science than manned missions. Measurement counts as manned when there is Kerbal onboard during the measurement. Doing manned measurement after you successfully delivered unmanned will get you only the manned surplus.

I like the one-shot-ness part. It's not like I can't send twenty transmissions, but sending twenty return missions to squeeze all the science from certain place sounds quite tedious.

[allmhuran]

Yes, I saw your suggestion and several others along similar lines, and they're all better than the 0.22 mechanics. From what I can see I think they're probably pretty close to what Squad is looking like doing with 0.23.

As you say, the big difference here is that I am suggesting the complete elimination of repeated experiments. This goes hand in hand with the removal of transmission loss, since each directly suggests the other because:

- If there is transmission loss, repeating experiments would have to be possible

- If repeating experiments is possible, then there must be some kind of diminishing return down to a capped total to prevent exploitation.

I think removing both mechanics simplifies things immensely and removes the monotony. I also think samples being returned and data being transmitted is a lot more intuitive than the current mechanic. Put those two changes together and you pretty much necessarily end up with what I have suggested.

[Kasuha]

I think removing both mechanics simplifies things immensely and removes the monotony. I also think samples being returned and data being transmitted is a lot more intuitive than the current mechanic. Put those two changes together and you pretty much necessarily end up with what I have suggested.

Data, if recorded using analog devices, can provide more information when returned than when transmitted. Of course it could be expected that main value will lie in the transmit part and the delivery may only provide some additional details.

Samples can be investigated in place - color, smell, consistency and a lot of other physical properties can be conveyed via transmission. So there could still be some transmission value even though main deal of science data can likely be only achieved by delivery.

Don't forget we know a lot about Mars without a single sample being delivered from there.

Edited November 19, 2013 by Kasuha

[allmhuran]

Agreed, but the idea here is to make the game better, not necessarily to simulate what happens in the real world - unless doing so would make the game better.

Real world science is complex and quite monotonous. In this particular case I don't think the right approach is to emulate reality in the game as accurately as possible, but rather to come up with an enjoyable (and preferably intuitive) model.

Regarding the analysis of samples - yeh, I think the lab module is a good addition to 0.23 and I think I have covered the idea of transmitting some data about samples by making use of that new part to get the data out of the sample, so to speak.

[peadar1987]

I'd prefer more experiments, more science apparatus, more biomes, day/night variation, and far higher diminishing returns/one shot experiments. I think that would remove much of the "grinding" (at least from our end, it's probably an immense amount of work for Squad to make 20 or 30 different biomes for each planet, with descriptions of the results for maybe 15 or 20 different experiments performed in each one, seismic readings, radiation levels, magnetic field, soil composition, gravity, volcanism, atmosphere, etc.)

And also some science reward for a manned base or space station, perhaps a one-off science payment when you add a specific module to a designated space station core part. Different science modules could have different requirements, some need lots (and I mean LOTS) of power, some consume fuel, some need Kerbals to man them. These could be along the lines of "zero-G materials laboratory", "self-contained environment experiment", "ant colony", or whatever. I think the lack of a real point to space stations and bases is one of the major gripes most people have with science in .23

Edited November 19, 2013 by peadar1987

[allmhuran]

Yep, I agree that we need reasons to build space stations, permanent bases, etc. I put up another post with ideas for "reasons to do various types of stuff". If you can think of more good gameplay mechanics for space stations and bases that would be great.

[SunJumper]

I can see it being useful to have the sensors automatically stream science over long periods of time. For example, the accelerometer sitting on the surface for a few days recording the frequency of quakes. This therefore leads into mapping the bodies, whether image wise, topographically, gravitationally, thermally, etc. I suppose science would be based not on biomes but on arbitrary subdivisions, you'd get science once per division but would also walk away with a few cool looking maps per body. (Then again, mapping bodies is handled by three mods, and the maps are only topographic/biome-wise/resource-wise).

Also, the materials bay might also count as a sample. I'd imagine it containing a few sub-samples - plastic,metal,etc.

Edited November 19, 2013 by SunJumper

[JumpsterG]

I had to think about this for a while before coming to good conclusions... I think squad is going in at least of few of the directions I mention here.

Since there are various different flavor texts for doing experiments in the same area, it's implied that you'll do it a few times (and should still be fun). This favors the diminishing returns strategy by rewarding you big time for the first successful mission and slowly discouraging repeats. For return-to-Kerbin missions, I have found no problem with the system as it is. You get full value, and it only takes a handful of missions (or duplicate science parts on the same mission) to collect everything from an area.

The problems with the system come from the transmission mechanics. There are 2 main problems:

-Transmission loss results in players repeating experiments in the same area many more times than is fun

-A power generating craft can do science infinitely and obsoletes return-to-Kerbin strategies.

The time sink involved with transmitted experiments is annoying. It forces you to repeat the same experiment, read the same flavor text, and sit through transmission/recharge cycles many times. Even if you split this up over multiple flights (which you don't need to, which I'll cover next), you will still run some experiments 20+ times to collect all science possible. Simply giving maximum value per experiment would reduce the amount of time seeing the same science report and would be simpler.

Power generating craft are capable of draining all science from multiple areas during a single mission while reusing a single science part.

This completely obsoletes the return-to-kerbin method of science collection which requires multiple science parts to carry experiment results and must make a return trip to Kerbin after collecting. A limit to the number of experiments that can be run per part would encourage the planning of multiple missions and/or require craft to be designed to carry additional science instruments in order to do more (the same as a return-to-Kerbin mission, but still with more capability to repeat experiments and without needing the return). To increase the usefullness of return-to-Kerbin missions to match the flexibility of transmission-based missions, some experiments should only be rewarded for recovery. This could be extended by the late game Kerballed lab module which might take over the recovery of data. This way, big rewards would remain for the more difficult return missions or those involving heavy space station or surface labs.

TLDR

My suggestions:

-Keep diminishing returns (We like some repeats and multiple flavor text per area)

-Drop transmission loss (We don't like 20 repeats per experiment per area)

-limit # experiments per science part (Make 1-shot missions to collect all science from one or more areas challenging)

-Differentiate between transmission-capable and lab-processed experiments (Maintain a use for return trips or heavy, Kerballed Lab-modules)

[Queeg]

I too would like to see a more intuitive solution, rather than a "gamey" one.

Diminishing returns seem to be solving a gaming problem.

Transmission quotas... same thing.

Plus both add complexity that just ends up being confusing to new players (or a least confused THIS new player). Maybe the Science Journal/Log will help. Maybe the science redesign will help. I'll wait and see.

But I really don't fancy seeing science turn into "kill 10 rats" type questing. Neither do I fancy needing to do advanced calculus to figure out if I should be killing 5 rats or 15.

I like the idea of data transmission and physical return. I like the idea that thermometer readings and crew reports are data and goo samples and soil samples are physical things that need returning to Kerbin (or a space station with the necessary science equipment) for analysis. That seems legit to me. It fits in with my view of how experiments are done. I'd even accept transmission errors, where reports are corrupted during transmission and needed repeating, with a greater chance of corruption the more links there are in the communications chain. But a flat percentage penalty doesn't feel fun, diminishing returns doubly so.

Fingers crossed for a "keep it simple" solution.

[whistlehead]

While I quite like what you've suggested, I'm going to stick my neck out and say that diminishing returns are to some extent a Good Thing - the reason being that in real life data from an experiment is only considered reliable if it can be repeated, this being the entire basis of the scientific method. Fi the game were to be used as a means of encouraging scientific though in younger children/teenagers, for example, then I'd like to see some element of the true scientific method in this magic resource we call Science.

That doesn't mean the current system works. I just don't think it's sensible for experiments to be a one-shot deal, as this is simply not how science truly works.

Regards,

Chris

[allmhuran]

Yeah, a lot of people seem to bring up this objection in some form. "Real life science requires repetition".

Here's the thing: That's really boring as a gameplay mechanic. Hell, it's really boring in real life too.

Here's the other thing: If we REALLY wanted science to be like real life, we'd get zero points for performing an experiment the first time. Surely everyone knows experimental results are of no value at all until you repeat them many times.

Here's the final thing: Nobody seems to care that we don't have to fill out flight plans, or wait an hour for our Kerbals to perform a checklist before going on EVA, or have cranes and platforms to build our rockets instead of having them just float there magically in the VAB, because these would be terrible for gameplay.

[whistlehead]

we'd get zero points for performing an experiment the first time

I did actually think of that immediately after posting...

Nonetheless, diminishing returns are still a positive in my mind - maybe a faster rate of reduction though.

[allmhuran]

I'm dead against the idea of diminishing returns in any form, so I'm curious what positives you think it adds to the gameplay. Being "more like real life" does not in itself make the game more fun, so I don't think that's an appropriate response.

[Unfawkable]

I actually have nothing to add, agreed on all points. One addition might be that it would require more energy to transmit data from planets that are further away, as they should give more science points as well.

[Tw1]

That doesn't mean the current system works. I just don't think it's sensible for experiments to be a one-shot deal, as this is simply not how science truly works.

Regards,

Chris

Agreed. I like to do the tests over and over again, even in spots I've been before, just incase they might find something new. (I know they won't, it's a roleplay thing.)

Though the "Science points" may diminish, I'd like a secondary system where you collect actual data- rock composition, atmo comp, even if the substances named are fictionalised- It would make it more interesting.

For example, regolith composition could vary based on distance from certain coordinates, and if they went down to a few decimal places, you'd only have to drive so far to notice the data changing. Maybe they could be used to lead you somewhere.

This wouldn't be for more points, this would make sure there were things to do on a long term mission, and after you've exhausted all the science points. Because really, science never ends. There's still so much we don't know about our own planet, and we've been here for ages.

[HeadHunter67]

diminishing returns are to some extent a Good Thing - the reason being that in real life data from an experiment is only considered reliable if it can be repeated, this being the entire basis of the scientific method.

I agree. And there's no reason why everything that can be determined about an area of study can, should, or must be contained within a single transmission. If that were the case, there'd only ever be one book, one YouTube video, one forum post on any given topic. The fact of the matter is, continued study should continue to yield new insights - even if the increase to the overall body of knowledge is incrementally smaller each time.

In short, diminishing returns from repeated transmission is the way information works. The only problem in the game is that it's possible to completely exhaust the available Science on a topic in this fashion, to the extent that there's no point in returning with the data.

And for those who feel that a sample must be returned to gain any scientific value, I'll point you to a little planet called Mars - whose surface we have been studying and learning about for decades, in spite of the fact that no surface samples have yet ever been returned for further examination.

[allmhuran]

Pointing to how things work in reality is rarely sufficient justification for making them work that way in a game.

If you can come up with a way of repeating experiments that's *fun*, I have no objection. But all of the proposed mechanics that I have seen end up with players sitting around waiting and click spamming. That's too high a cost.

Besides, "real science" involves a heck of a lot more than just repetition. So the claim that we should have repetition because that makes it "like real science" seems to be accepting a very high price (the monotony of repetition in the game) for a very low benefit (still utterly unrealistic)

Here's a more logically structured presentation of my thought process:

Given:

• Transmission is easier than return, so must be penalized.
  
• Transmission loss implies repetition.
  
• Repetition implies diminishing returns.
  

The problem is that diminishing returns implies more repetition. Therefore diminishing returns is not a solution to the repetition problem, and repetition is not a solution to the transmission problem.

The core of the solution I am proposing is therefore the natural solution to this problem: The elimination of transmission loss and diminishing returns entirely, by the elimination of repetition.

But this leaves us with the problem that transmission is easier than return. With no transmission loss, no repetition and no diminishing returns, there would be no penalty for transmitting.

My suggestion is therefore that this is more sensibly addressed by splitting the science into samples and data, allowing data to be transmitted (for 100% of its value, since there is no other possibility), not allowing samples to be transmitted, and making samples much more valuable than data.

Everything else is simply an extension of this.

Now, several posters have missed the core of this idea and instead have tried to defend the repetition mechanic. The most common retort has been "science in real life requires repetition, so repetition should be in the game".

I defeat this claim threefold:

1. Realism in and of itself is not sufficient justification to add something to the game. If it were, then we would also be suggesting the addition of all kinds of things that are no fun and which clearly nobody actually wants added. So this claim is defeated by reductio.
   
2. Realism in this specific case is not addressed merely by the addition of repetition, since even with repetition KSP science is still utterly unrealistic, and we have gained nothing at the cost of having less enjoyable gameplay.
   
3. The first priority must be making the game fun because this is the very essence of a game. Monotonous repetition is not fun. KSP science involves monotonous repetition. Therefore we should eliminate it.
   

Edited November 20, 2013 by allmhuran

[Idobox]

About the repetition thing:

If I measure the temperature 3 times in one spot, or with 3 thermometers at the same time, I still have only one data point. Sure, I am now more sure my thermometer is not broken or badly calibrated, but that's it.

Now, if I measure temperature every hour, or every km, I get more data points.

If the developers really want to have a repetition mechanism, there are other means:

-give science points only if the experiment is repeated far away. Biomes already do that, but with low precision. A sort of mapping, with a big array of booleans to store whether an experiment has been done in one place or not might be an idea.

-upgrade science parts. The first thermometer could have a 10K precision, the second one 1K, and the third one .1K, giving you a reason to repeat the experiment once you have unlocked the new part. It could feel pretty grindy though.

-special experiments. Once missions are implemented, you could have an "experiment part", always the same but with different text, that needs to be sent somewhere and used n times, depending on the mission you might have to bring it back or leave it there. Once the mission is over, you would get new missions to complete. Like mission 1, send these ants into orbit and watch them. Mission 2, send this ants into orbit, and then spin the craft really fast. Mission 3, observe these ants in LKO, HKO, Mun orbit and on the surface of the Mun. You now have a reason a good reason to do the same experiment half a dozen times.

That being said, removing completely the diminishing returns mechanism means you will have no reason to go back to body X once you went there once with all the science parts.

A good way to limit that is biomes. Another is to keep the dimishing return mechanism, but allow only one transmission by mission by part type and by biome. There would be no point in putting 3 goo canisters on your first rocket to reach orbit if you could only reap science once, but there would still be a reason to put a goo canister on your 4th rocket to reach orbit. The risk would be to have players forced to repeat the same mission a few times to get all the science.

Another option Squad could have chosen but didn't was to go the old Japanese RPG way: killing a slime (or measuring temperature at KSC) gives you 10 points every time, but you need 7 000 000 points to reach level 62, so you rather try to kill a golem (or measure gravitrons on Moho) for 5 000 000 points. I think it's good thing they didn't, as we would have the option to measure temperature 700 000 times.

[Corona688]

It's not black and white. There's situations you can't return samples -- try getting anything back from Jool's core*. Transmission should be worth something, but shouldn't be as farmable as it is.

Also... It's possible to get stuck. I've always worried that could happen in career mode, and the ability to make small amounts of progress helps. If you've made bad choices in the tech tree it's hard to earn more science, but not impossible. They could make creative designs and do creative missions... or just farm. In practice it becomes a bit of both. I did a lot of Kerbin exploration and found some neat things, but when you're 1 point away from earning a vital part and out of biomes to explore...

Also... Making it black and white would also make research a lot less interesting. Jool is a good example actually. How to transmit enough data from Jool's atmosphere before your batteries die and/or your probe implodes? Make it bigger -- more batteries, faster transmitter. Make it bigger yet, multiple transmitters and sensors. Or make up the difference with lots of cheap probes. Or give it a rocket to generate power and slow its descent. Or more chutes and a horde of pasted-on solars. Or strap on some of those heavy, weak, but endless RTG's. Or arrive in an orbit which aerobrakes repeatedly, buying time to transmit. Or lots of things I haven't even thought of.

Partial progress through transmits lets you fart around, make mistakes, have fun, blow things up, struggle and still make (slower) progress. I think this is why they made research the way they did. We're good enough at the game to have science down to a science, know all the parts we're going to get, plan and breeze our way through the tech tree, planning missions around points, not curiosity. In short it's tedious because we are.

*Actually that sounds interesting now. Let me think on that.

Edited November 20, 2013 by Corona688

[Corona688]

That being said, removing completely the diminishing returns mechanism means you will have no reason to go back to body X once you went there once with all the science parts.

There are quite a few things to find on the moon. Craters, trenches, plateaus with their own readings to report. Once you go beyond that, things get simpler... I spent a long time on Minmus trying to get different samples from the 'seas' and the heights -- either I picked all the wrong spots or there's no difference.

Encouraging exploration is always a good thing.

[Kimberly]

I would like it if science value went in a high-low-high-low pattern. Your first data point is worth a lot, then they get steadily less valuable. But once you collect a lot of data points, you are able to use them to draw new conclusions and their value goes up--you are rewarded with new flavor text and possibly a display of your data, such as atmospheric composition, weather patterns, whatever is appropriate. After your discovery, you can continue to do research if you want, but its value drops off sharply.

In this way, you don't need to repeat yourself, but there's a reward for doing so. And the extra flavor means it's not monotonous.

[Jim DiGriz]

I can see it being useful to have the sensors automatically stream science over long periods of time. For example, the accelerometer sitting on the surface for a few days recording the frequency of quakes.

Yeah, big +1 on this.

I think you should get big immediate payoffs for bringing samples back, and that there should be a continuous trickle of passive science collection from bases and probes (that are appropriately equipped). The goo and the science container should probably only be for sample collection -- the other instruments should only be for passive. Sample collection should probably be one-shot-per-biome/level so you don't have to grind through sample collection. EVA and crew reports should also be one-shot and not grind, maybe you allow crew reports to transmit back, while EVAs they need to return or something -- well, not sure that makes sense... But, for passive science collection an unmanned probe should really trickle slowly, while a manned base with some really heavy piece of science-oriented gear should trickle a lot faster. Presence of Kerbals doing the collection should be a big bonus over unmanned missions.

[allmhuran]

I would like it if science value went in a high-low-high-low pattern. Your first data point is worth a lot, then they get steadily less valuable. But once you collect a lot of data points, you are able to use them to draw new conclusions and their value goes up--you are rewarded with new flavor text and possibly a display of your data, such as atmospheric composition, weather patterns, whatever is appropriate. After your discovery, you can continue to do research if you want, but its value drops off sharply.

In this way, you don't need to repeat yourself, but there's a reward for doing so. And the extra flavor means it's not monotonous.

It's cetainly an idea that could be incorporated somewhere... but it terms of what the player actually does, it still leaves them doing nothing but spamming the "do science" action group button, which is still tedious.

[Corona688]

Again, its only tedious if you play it that way. There's ways to make it more efficient.