A more intuitive tech tree

[Superfluous J]

I don't like that. It seems contrived for no purpose other than restricting what you can do.

Okay then put all the probe, manned, and plane parts in the initial node and then just branch off normally from there.

Also, "contrived for no purpose other than restricting what you can do" is a good description of the entire tech tree mechanic.

[CaptainKipard]

Okay then put all the probe, manned, and plane parts in the initial node and then just branch off normally from there.

Or you could have a compromise between the two: Have them all in separate nodes and have them cost a normal amount lol.

Also, "contrived for no purpose other than restricting what you can do" is a good description of the entire tech tree mechanic.

As it is now, yes, but it's meant to be an abstraction of technological progress, so there's a reason for having it, just not in the current form.

[Superfluous J]

Or you could have a compromise between the two: Have them all in separate nodes and have them cost a normal amount lol.

Yeah but then you get back to the "what do you put in the starting node" question.

For what it's worth, the tech tree I started fiddling with had the starting node the same as the current one, then had a "probes" tree (starting with stayputnik and going up slowly through each probe type) a "planes" tree (with a starting package about as fleshed out as the starting rocket package, ie a cockpit, single wing part, control surface, tail fin, engine, intake, fuel, and landing gear), and a few more trees (structural, fuel, engines, electronics, yadda yadda). I lost interest when there wasn't a simple program with drag and drop support to make the tree but hey, I thought it had promise.

[Baythan]

I don't see why there has to be a starting node with anything in it to begin with. Either have an option when you set up the career game where you select one of three starting trees, or you start career mode with a science/funds budget that lets you buy parts.

The buying parts option would work best if fewer parts were in each node and each node cost less, but there would be MANY more nodes. The option for mods to add nodes attached to pre-existing nodes would also be great. Of course, this makes the tech tree a lot more complex and would probably mean an extensive re-write of that section of the game... but that's pretty much what's needed in the end I believe.

Again, the overall consensus is: More choice when it comes to advancing in career mode.

[Bill Phil]

A tech web is probably a much better implementation of technological progress. You can have a certain number of categories that overlap in cool ways. The first node should have a Mk1 Pod, the Stayputnik Mk1, and a collection of parts similar to the starting tech currently in use. Although the RT-10 should be replaced in that node by a smaller 0.625 meter SRB that's about a tone in mass, and based off of the Delta IIs SRBs. The RT-10 should come later on. Maybe in the propulsion technology category of the tech web. Or something like that.

[Baythan]

A tech web is probably a much better implementation of technological progress. You can have a certain number of categories that overlap in cool ways. The first node should have a Mk1 Pod, the Stayputnik Mk1, and a collection of parts similar to the starting tech currently in use. Although the RT-10 should be replaced in that node by a smaller 0.625 meter SRB that's about a tone in mass, and based off of the Delta IIs SRBs. The RT-10 should come later on. Maybe in the propulsion technology category of the tech web. Or something like that.

But what about plane parts? I don't want to start with manned/unmanned rockets.

[Kegereneku]

But what about plane parts? I don't want to start with manned/unmanned rockets.

I don't see a problem with starting with plane because (1) that's realist, (2) there's a lot of biomes on Kerbin.

Once the tech-tree get reworked -hopefully toward a more choice based architecture- it should be more than balanced to allow plane without going straight into spaceplane.

[CaptainKipard]

I think the Start node should only have a few basic or small structural parts that are useful for all or most types of craft, like nacelle pylons or small girders.

Everything else should branch off, including plane parts. The player should chose.

[Baythan]

I think the Start node should only have a few basic or small structural parts that are useful for all or most types of craft, like nacelle pylons or small girders.

Everything else should branch off, including plane parts. The player should chose.

Agreed. There are a number of parts that are useful for all craft types that currently end up 3-5 tiers into the tech tree, like the cubic octagonal strut, small hardpoint and all nosecones, any of which could be put much earlier. If we get that Drag model revamp soon we'll really want those nosecones early and possibly some fairings introduced into stock.

[sherkaner]

I've been playing with taking the current parts and actually dropping them into a tech tree like we're talking about (I'll probably post that here soon). Among other things I'm realizing as I work through it, I think a "starting node" really may not be necessary. I think the main purpose of starting node parts is so that you don't have players making bad choices with their starting science that leave them unable to do anything, but in practice I'm finding that with just a few logical dependencies you end up with a set of initial choices of parts that are *very* flexible in letting the player choose what they find interesting, but don't easily let you "waste" science on stuff that doesn't work.

Alternatively, it also seems like you might just give all of the entry point parts to a player off the bat to encourage them to play around without feeling like they have to make choices they can't undo. I don't think it would be so bad to give them the parts to build a very very simple rocket, plane, and rover automatically.

Plus I'm just overall finding that (surprisingly to me) even with the current stock parts you can create a pretty satisfying tree that feels both realistic (ie. the dependencies feel like meaningful progress on a part type rather than arbitrary) and seems like it would balance out gameplay wise. I'm liking it so much, I may take a shot of putting it into a config file for the custom TechManager mod so I can play a career with it. I'll definitely be interested in feedback from you guys first though.

Edited December 24, 2014 by sherkaner

[CaptainKipard]

I'll definitely be interested in feedback from you guys first though.

If you're using the default nodes and just moving the parts then I'd be interested in trying. All my mod parts will obviously be in the wrong place, but I'll just have to suffer through it.

[sherkaner]

If you're using the default nodes and just moving the parts then I'd be interested in trying. All my mod parts will obviously be in the wrong place, but I'll just have to suffer through it.

Yeah, I need to look at this mod more and figure out how much I can define. I'm really hoping to do away with the default nodes and define nodes that almost always contain one part. Stock parts only obviously. Not sure if the mod is that flexible though. Either way I'll post my tech tree diagram organized that way so you guys can critique.

[sherkaner]

Ah, nice, it looks like the .cfg files that KSP/TechManager use are fully flexible for defining nodes as well as parts within, so totally doable for me to create the tree we're talking about. Might even be able to have it in "explosion" format. It'll take me some time though. I'll get the diagram up first later this evening.

[Baythan]

I'd love to have a look at what you work out. Almost anything that anyone would find enjoyable would be better than what we have now.

[sherkaner]

Okay, here's the tech tree I've been working on incorporating all stock parts. It's in linear format, but of course it could be wrapped into an explosion format. Of course I think we're all expecting Squad to modify, rebalance, and possibly add parts during beta, but I was surprised how well this worked out even using today's parts -- vastly better than the grouped tree we have today for providing flexibility and a sense of guiding a "real" R&D program, while still providing challenges and tough choices:

Note: New version of this diagram now posted later in this thread here.

Yeah, it's huge.

The guiding principle here is the "per part" nodes concept we've been discussing where each part type essentially has its own branch in the tree with a sequence of branching nodes for more and more advanced versions. I've tried to stick to that philosophy pretty closely, but here are a few additional background notes:

• If you look at the first parts in each category on the left, I think those provide a very nice cross-section of basic parts that will provide players with many choices but not be too "overpowered" or get the player in trouble researching useless parts. Basically with those you can create a basic rocket, plane, or rover and do a little science. I think these might even be parts to just give the player outright.
  
• You'll see that I sometimes have multiple inputs to a single node -- these are meant to be multiple options of getting to that part, NOT multiple prerequisites. I did this in places where there were "realistic" feeling different ways you'd get to a certain technology by different paths, and also to add flexibility in places where -- for example -- both spaceplanes and rockets could use a part and I wanted to avoid forcing somebody to dig through a tree branch they'd otherwise be uninterested in.
  
• While I generally stuck to "one part per node", I do group a few parts together that were so similar that making them separate nodes seemed ridiculous.
  
• I basically punted on the wings section, describing in general terms what I think should be at the different R&D levels. These were the only parts that seem so "broken" currently that organizing them was almost nonsensical. The pre-procedural wing/fin parts just make no sense, but I wager all of that is about to change soon anyway.
  
• I could use some input on the structural (i-beams, girders, etc.) parts. I haven't had much opportunity to use those yet, so I'm not sure how well my choices play out. But generally I think these parts should be available sooner rather than later.
  

Anyway, looking forward to hearing what people think, and maybe after I make some revisions I'll work on getting this into a .cfg for TechManager so we can play it. Personally I'm super-excited about playing a career with something like this tech tree. I can imagine a lot of very different-feeling programs being played through this tree to a high level with very different sets of technologies unlocked. And as an engineer, this to me feels like something much closer to the organic way that real technology developments evolve and arrive at solutions from different angles.

Edited December 27, 2014 by sherkaner
Merry Christmas

[Baythan]

Okay, here's the tech tree I've been working on [snip]

I like it!

I might move the mobile processing lab into the science branch rather than the habitat/command pod branch. Also, I think the QBE probe might be too early, but I don't know what pilot skill level it has programmed in.

Next would be balancing the research cost of each node/part.

[sherkaner]

I like it!

I might move the mobile processing lab into the science branch rather than the habitat/command pod branch. Also, I think the QBE probe might be too early, but I don't know what pilot skill level it has programmed in.

I debated on the processing lab. I kept it in habitat just because it does contain Kerbals and so purely R&D wise I thought of the main development of the lab (as opposed to its usage) would be more similar to pods and landing cans.

And yeah, I need to look into the QBE -- good point. I was making the assumption that it was very basic (and cheap), but I'm actually not sure with 0.90. I probably need to do some sandbox play with the new probe cores in general before I settle that branch.

Next would be balancing the research cost of each node/part.

Yeah, that's going to be interesting. Clearly each node is going to have to be much cheaper than current (since you're only getting one part generally), but because the player can be very selective, you can probably make the parts somewhat expensive. I'm open to suggestions, if somebody with more career experience wanted to mark up my tree with costs...

[Pthigrivi]

Just a little off-topic chime-in, but has anyone explained why the small hardpoint, radial attachment point, and 1.25m quad coupler are so crazy expensive? I mean they're handy and I like them but they're like 3x the cost of the average structural part. The 2.5m quad coupler is like 40% the price of its little brother, the Hubmax is 70% the cost of radial attachment point, and the small hardpoint seems just crazy overpriced.

That's it, carry on!

Edited December 25, 2014 by Pthigrivi

[Kegereneku]

Okay, here's the tech tree

Nice work, now that's something I would actually be interested of playing with.

Sincerely, I wouldn't worry much about order and balance, not only people will disagree on a few point but news parts with new interaction, and another tiers not-yet-added (if I remember) make it subject to big change.

A few remark :

• You'll see that I sometimes have multiple inputs to a single node -- these are meant to be multiple options of getting to that part, NOT multiple prerequisites. I did this in places where there were "realistic" feeling different ways you'd get to a certain technology by different paths, and also to add flexibility in places where -- for example -- both spaceplanes and rockets could use a part and I wanted to avoid forcing somebody to dig through a tree branch they'd otherwise be uninterested in.
  

Myself I would actually support multiple prerequisites for a few carefully chosen part.

Yes, that's surprising coming from someone who loath the bundles we have now, but a few parts isn't a big cost for the following reasons :

- Encouraging new players to try out new part they wouldn't think to unlock first.

- As a balance mechanism (limiting minimaxing even though R&D tier take care of that at 80%)

- Minimal verisimilitude (less people complaining about unrealistic tech order.)

• I basically punted on the wings section, describing in general terms what I think should be at the different R&D levels. These were the only parts that seem so "broken" currently that organizing them was almost nonsensical. The pre-procedural wing/fin parts just make no sense, but I wager all of that is about to change soon anyway.
  
• I could use some input on the structural (i-beams, girders, etc.) parts. I haven't had much opportunity to use those yet, so I'm not sure how well my choices play out. But generally I think these parts should be available sooner rather than later.
  

I thing the way we could make sense of wing parts would be to put the wing form that are "the most practical" (in term of achieving exactly the wing profile you want) on later tier.

For example, it would be impossible without clipping (which can intersect inside cargo-bay) to obtain more than 45% Delta wing.

As for structural part, from my opinion they should be nearly all unlockable in Tier 1 except for those based upon 2.5m, 3.75m, mk2 and mk3 structure (which would be prerequisite)

My 2cents

Edited December 25, 2014 by Kegereneku

[sherkaner]

Myself I would actually support multiple prerequisites for a few carefully chosen part.

Yes, that's surprising coming from someone who loath the bundles we have now, but a few parts isn't a big cost for the following reasons :

- Encouraging new players to try out new part they wouldn't think to unlock first.

- As a balance mechanism (limiting minimaxing even though R&D tier take care of that at 80%)

- Minimal verisimilitude (less people complaining about unrealistic tech order.)

You make some good points, and I there are a couple of places where I sort of wanted to have multiple prerequisites (the LFB KR-1 was one in particular). But I just felt like there weren't enough parts to justify forcing people into building out whole branches just for a prerequisite. I think if Squad were to add a lot more endgame parts, that's where I'd be most in favor of it to make people work a little harder for really awesome high-tech stuff.

But it might be interesting to try. Basically anywhere that I have multiple input lines to a part, it would be easy enough to create a version of a tech tree that treats those as prerequisites rather than options. I might try both.

I thing the way we could make sense of wing parts would be to put the wing form that are "the most practical" (in term of achieving exactly the wing profile you want) on later tier.

For example, it would be impossible without clipping (which can intersect inside cargo-bay) to obtain more than 45% Delta wing.

True, I might be able to split up the procedural parts a bit that way, within R&D level 2, good idea.

The non-procedure stuff (that I'm lumped into R&D level 1) is where I'm just totally baffled. But I think those are also probably the ones likely to change a lot -- especially with the mention that Squad's looking at aerodynamics soon. So just for the sake of playtesting this tree for now, I'll probably just throw all of those into one or two nodes and call it a day.

As for structural part, from my opinion they should be nearly all unlockable in Tier 1 except for those based upon 2.5m, 3.75m, mk2 and mk3 structure (which would be prerequisite)

My 2cents

That was my feeling mostly also, and pretty much what I have in the diagram I think. Any in particular there that you'd disagree with?

I'll definitely start trying to get this dumped into a tech tree .cfg file in the coming days.

Keep the input coming!

[CaptainKipard]

I like what I'm seeing a lot. And the diagram itself is really elegant.

I looked at it in detail and I have a bunch of question and a bunch of suggestions but they'll have to wait a bit.

One thing I wanted to ask you is whether you have all those part images saved as separate files, and if you could upload them for those who'd like to try designing their own trees.

[klgraham1013]

Okay, here's the tech tree I've been working on incorporating all stock parts. It's in linear format, but of course it could be wrapped into an explosion format. Of course I think we're all expecting Squad to modify, rebalance, and possibly add parts during beta, but I was surprised how well this worked out even using today's parts -- vastly better than the grouped tree we have today for providing flexibility and a sense of guiding a "real" R&D program, while still providing challenges and tough choices:

http://i.imgur.com/hlrFxLT.png

Yeah, it's huge.

The guiding principle here is the "per part" nodes concept we've been discussing where each part type essentially has its own branch in the tree with a sequence of branching nodes for more and more advanced versions. I've tried to stick to that philosophy pretty closely, but here are a few additional background notes:

• If you look at the first parts in each category on the left, I think those provide a very nice cross-section of basic parts that will provide players with many choices but not be too "overpowered" or get the player in trouble researching useless parts. Basically with those you can create a basic rocket, plane, or rover and do a little science. I think these might even be parts to just give the player outright.
  
• You'll see that I sometimes have multiple inputs to a single node -- these are meant to be multiple options of getting to that part, NOT multiple prerequisites. I did this in places where there were "realistic" feeling different ways you'd get to a certain technology by different paths, and also to add flexibility in places where -- for example -- both spaceplanes and rockets could use a part and I wanted to avoid forcing somebody to dig through a tree branch they'd otherwise be uninterested in.
  
• While I generally stuck to "one part per node", I do group a few parts together that were so similar that making them separate nodes seemed ridiculous.
  
• I basically punted on the wings section, describing in general terms what I think should be at the different R&D levels. These were the only parts that seem so "broken" currently that organizing them was almost nonsensical. The pre-procedural wing/fin parts just make no sense, but I wager all of that is about to change soon anyway.
  
• I could use some input on the structural (i-beams, girders, etc.) parts. I haven't had much opportunity to use those yet, so I'm not sure how well my choices play out. But generally I think these parts should be available sooner rather than later.
  

Anyway, looking forward to hearing what people think, and maybe after I make some revisions I'll work on getting this into a .cfg for TechManager so we can play it. Personally I'm super-excited about playing a career with something like this tech tree. I can imagine a lot of very different-feeling programs being played through this tree to a high level with very different sets of technologies unlocked. And as an engineer, this to me feels like something much closer to the organic way that real technology developments evolve and arrive at solutions from different angles.

This is really good. If this were stock, I'd definitely play science career.

[Kegereneku]

That was my feeling mostly also, and pretty much what I have in the diagram I think. Any in particular there that you'd disagree with?

I'll definitely start trying to get this dumped into a tech tree .cfg file in the coming days.

Keep the input coming!

I as said, I would take your picture as a concept demonstration rather than a route plan to polish.

It's expected to go through many change, Engines stats are going to change with the upgrade of the Aerodynamic model, new parts are to be expected, all might create new dynamic susceptible to ask for a reorder. Also one token guy might consider the Dockport .jr should come first as it is technologically less advanced...etc

Plus, in the end it's SQUAD who will have the last words.

Edited December 26, 2014 by Kegereneku
it really needed correction

[CaptainKipard]

OK Questions first:

1. Why did you make the stability enhancer dependent on the radial decoupler?

2. Why place the mini decoupler and separator after the medium?

3. Same question for the clampotron junior.

4. Same question for Oscar fuel tank.

5. I think I already know the answer but why make the klaw dependent on docking ports? It's a modified docking port out of laziness on Squads part (or time constraints) more than anything else.

6. What does the line connecting the "Mk2 Cockpit" and "Mk2 Liquid fuel fuselage short" mean?

7. Why is the structural intake last in that branch?

8. Why are the radial RCS tanks so late?

9. Why are the Xenon tanks dependent on all the RCS tanks before them?

10. Why make RCS thrusters dependent on the the little orange LFO engine?

11. Why make the O-10 and Vernor dependent like that and not the other way around? I think you went by function, rather than technology.

12. Why make the ion engine technically dependent on a battery? I know it's practically dependent, but you could do it the way you did the RTG. Available late, with no dependency.

13. Why is the gravioli detector not a part of the sensor branch?

You can probably tell a lot of these are veiled suggestions, but I'll go through my exact thoughts later when you answer.

[sherkaner]

I like what I'm seeing a lot. And the diagram itself is really elegant.

I looked at it in detail and I have a bunch of question and a bunch of suggestions but they'll have to wait a bit.

One thing I wanted to ask you is whether you have all those part images saved as separate files, and if you could upload them for those who'd like to try designing their own trees.

Ah, yeah, I do actually. They're basically just screenshots from the KSP wiki, but for convenience here you go:

http://www.filedropper.com/kspparts