ShadowZone - What Kerbal Space Program 2 HAS TO Avoid

[TROPtastic]

@ShadowZone released a new video today on a list of things that KSP2 has to avoid in order for it to be a game with more success and player engagement than KSP1.

I won't make a full summary here (the video is worth watching), but the list of items covers mission-ending bugs related to landing gear, wheels, docking, and the usability of the orbital maneuver planner. I fully agree with the list, and I'd go further to say that solving these problems is necessary but not sufficient to make a game that is more successful than KSP1. We know that the devs are working on some other aspects to make KSP2 more accessible (the cool looking tutorial animations), and hopefully there will be an in-game incentive to leave the Kerbin-Mun-Minmus system that was lacking in KSP1.

 

[Bej Kerman]

This resonates with me SO MUCH. KSP is full, full of these kinds of seemingly minor bugs that can brick your missions (like the docking port resave bug) and just general jankyness (this is not how wheels work!!!), and I am exceedingly enthusiastic about being able to ditch this game for good. 

Edited February 10, 2023 by Bej Kerman

[regex]

Quite frankly what KSP2 needs to avoid is just being an updated KSP1. New (main vehicle-focused) gameplay, new building methods (for vehicles), new things to do (that don't involve right-click then left-click).

[darthgently]

Maybe too late for KSP2, but a fully mod accessible UI layer would be great.  So the entire game interface could be defined  by a mod

[TROPtastic]

1 hour ago, Bej Kerman said:

This resonates with me SO MUCH. KSP is full, full of these kinds of seemingly minor bugs that can brick your missions (like the docking port resave bug) and just general jankyness (this is not how wheels work!!!), and I am exceedingly enthusiastic about being able to ditch this game for good. 

Completely agree. I think ShadowZone said it best when he said "I don't want to be a Kraken wrangler, I want to be able to enjoy myself without creating constant manual backups of my save games because I cannot trust the game to work as intended." Making KSP2 from the ground up should hopefully solve all the annoying bugs in the KSP1 codebase.

[Bej Kerman]

1 hour ago, regex said:

Quite frankly what KSP2 needs to avoid is just being an updated KSP1

And it isn't

[Scarecrow71]

I love SZ's videos.  Good information here, even if he is merely stating what is already known and what he hopes for.  Now I gotta go watch the Invictus vids!

[K^2]

Wheels is one of these things that everyone keeps getting wrong, even though it's not actually that complicated. Just a bit unintuitive in places. Get someone who knows what they're doing to implement your game's wheels. Intercept, if you're looking for somebody, I can do a contract. :p Seriously, don't just hand it to someone enthusiastic but clueless and tell them to figure it out.

[Luriss]

Y'know, I think SZ forgot to mention the biggest problem with KSP 1 in his video; performance.

I'd argue the reason people don't go interplanetary is mainly due to the fact that in order to do so you have to build a 100 - 200 part ship which guarantees bad FPS, even on decent computers; the bugs are just icing on the cake.
Take me for example, I've been playing KSP since 2012 and the farthest I've ever gone is a one way trip to Duna, right about the time my career games really start to chug.

[The Aziz]

7 hours ago, Luriss said:

in order to do so you have to build a 100 - 200 part ship

Not really. Unless you aim for return from Eve in the same ship that gets there.

 

[Luriss]

5 minutes ago, The Aziz said:

Not really. Unless you aim for return from Eve in the same ship that gets there.

 

Yeah I was think more along those lines. A mothership with attached landers and such.

[The Aziz]

For Duna-and-back all you need is a 2.5m lander with Wolfhound at the bottom (or Poodle, don't remember the stats on the surface), about 2k dV, so Rockomax X200-16 or -32 is enough, docked to another 2.5m nuclear ship, with Jumbo tank. Launch separately, dock in LKO. 

So really the tricky parts are launching the big tank (nothing that can't be solved with 3.75 or 5m parts to keep the count low), and two dockings, one here, the other at Duna. Full craft that is then en route to Duna will have much less than 100 parts.

Other interplanetary missions can be solved by sending everything separately and meeting at destination. That's how I did my Moho mission, and returned using a small nuclear return vehicle.

Problems with the game really start to uncover if you decide to go full Kerbal like SZ with his invictus.

[darthgently]

The challenge I set is to recover or repurpose everything launched. 

I leverage Stage Recovery and parachutes for lower stages on Kerbin launches, do a lot of kOS scripting for rendezvous/docking to assemble pieces into a mission, and rely heavily on ISRU fuel from Minmus when venturing beyond the Kerbin system. 

The payload cost of Kerbin launched fuel really adds up if you don't ISRU off world.

Once you get cheap fuel, launching mostly empty tanks is easy from Kerbin and allows for fairly luxurious missions elsewhere.

I haven't fully figured out full reusability for Eve surface return missions, but I think I've seen where others have done it, not sure though

[Bej Kerman]

2 hours ago, The Aziz said:

For Duna-and-back all you need is a 2.5m lander with Wolfhound at the bottom (or Poodle, don't remember the stats on the surface), about 2k dV, so Rockomax X200-16 or -32 is enough, docked to another 2.5m nuclear ship, with Jumbo tank. Launch separately, dock in LKO. 

So really the tricky parts are launching the big tank (nothing that can't be solved with 3.75 or 5m parts to keep the count low), and two dockings, one here, the other at Duna. Full craft that is then en route to Duna will have much less than 100 parts.

Other interplanetary missions can be solved by sending everything separately and meeting at destination. That's how I did my Moho mission, and returned using a small nuclear return vehicle.

And is everyone into optimization? No, of course not. Some people aren't into doing rendezvouses all the time, others haven't even figured out how to dock. Point being, I see this argument all the time, as if a game like Kerbal Space Program doesn't need to be built to support reasonably sized rockets. It does.

2 hours ago, The Aziz said:

Problems with the game really start to uncover if you decide to go full Kerbal like SZ with his invictus

Black and white fallacy. Just because you're not hyper-optimizing everything doesn't mean you aren't building rockets large enough to put a city in orbit.

[The Aziz]

I was merely pointing out that you don't need a huge ship to go for your first interplanetary mission (and new player probably wouldn't build a WH40k flying cathedral anyway). So performance isn't the reason people stay in the Kerbin system. The learning process, or rather lack thereof, is. KSP2 seems to solve both learning (tutorials) and performance (the intent to construct colonies and interstellar vessels) issues so ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯

[Bej Kerman]

Just now, The Aziz said:

I was merely pointing out that you don't need a huge ship to go for your first interplanetary mission (and new player probably wouldn't build a WH40k flying cathedral anyway)

I'm pointing out that new players are going to be using a ship large enough to carry a transfer vehicle and lander to orbit, and that's where KSP begins to struggle.

1 minute ago, The Aziz said:

So performance isn't the reason people stay in the Kerbin system

Yes it is, unless you are implying that everyone owns a beast of a computer.

1 minute ago, The Aziz said:

The learning process, or rather lack thereof, is

The learning process is only one part of the problem. I know how to dock. I know how to build vessels. I'm still not bothered playing KSP, because unless the only thing I plan to do is to send a dinky little lander, plant a flag and go right back, the game will struggle. If I dare to do anything with an inkling of ambition, the yellow clock tedium begins. Players shouldn't be discouraged from sending anything bigger than a lander.

[Starhawk]

1 minute ago, Bej Kerman said:

8 minutes ago, The Aziz said:

So performance isn't the reason people stay in the Kerbin system

Yes it is, unless you are implying that everyone owns a beast of a computer.

I don't think this is correct.

I'm fairly sure that psychology and the orbital mechanics learning curve are the large factors keeping the majority of players from leaving Kerbin.

Happy landings!

[Bej Kerman]

Just now, Starhawk said:

I'm fairly sure that psychology and the orbital mechanics learning curve are the large factors keeping the majority of players from leaving Kerbin.

I'm not saying otherwise. I'm just saying that lag is also another factor and can also serve as a psychological deterrent 

I'll just leave this here because it's very relevant.

On 1/12/2023 at 1:17 PM, Master39 said:

The biggest frustration I had about KSP pre-2019 was when I saw a few clips of a youtuber sending this huge and complex colonization ship to Duna. It had dozens of containers filled with all the resources, crafts and equipment needed for this huge project.
Frustration because that youtuber clearly said multiple times that, despite him having a top of the line PC the framerate wasn't playable, what we saw in the video was speeded up to be viewable, and that it required plenty of quicksaves and backups because of things constantly bugging out.

That frustration of knowing that I would never be able to make missions like that, that I can't trust the game to be stable and in working order whenever I do anything simpler than a direct mission with a couple of crafts, that If I try to plan to reuse the same mother-ship for multiple missions I have to take into account that things start bugging out if you undock and redock the same crafts too many times.

I can't simply trust the game for anything more complex than planting a flag and then running home immediately and with the lowest part count possible.

I started hoping for a new studio to pick up the KSP idea and properly making a game out it a long time before they even announced KSP2.

And that's not to crap on KSP1, on Steam I have 1400 hours, but I know that's only a portion of the time I've spent on that game, I know the limits of a first game made by an indie studio that wasn't even a game development company, it's a good game for the tools they had.

Yes Aziz, you can just plant a flag and minimize part count, but that's boring.

[Pthigrivi]

These are all really good points and I really appreciate Shadowzones approach to all this. Having the game be better optimized and less buggy is important, but so is teaching players how to be more efficient. Both things can be true. Hopefully the tutorials help a bit with this, and down the road the science and resource systems may provide further incentives to make vessels as lean as they can be no matter how ambitious the plan. There is something else though that hasn't been mentioned here and that is KSP2 needs to give players better flight information. And I don't mean clogging up the main interface with lots of numbers, I mean graphical overlays that help players navigate and fly better. Some of this looks already to be in the works with much improved UI around maneuvers and burns, but there are 3 other categories we haven't heard much about yet that would be absolutely critical to helping players get past the interplanetary hump and start doing more complex things like establishing colonies:

1) An easy to use and understand Mission planner that lets players lay out their desired journey and provides them with dV totals and launch windows and lets them set alarms right there. Ideally the planner would be graphically simple and default to the next window, but offer drop down options to refine or alter launch dates using a porkchop plot. Even better would be the ability to save these mission plans to vessels or workspaces in the VAB so a player could look right there and see "Oh it looks like I'll need 430 dV to land on Dres but my lander only has 300... I'd better beef that up."

2) Better visual cues for precise landing much like what is found in Trajectories--visual targets and landing zone prediction, estimated difference between target and predicted landing point, trajectory factoring drag, perhaps even overheat prediction for reentry profiles. I'd also love to see some of the things Better Burn Time included like automatically showing time to impact/rendezvous and estimated burn time, estimated time to enter/exit an atmosphere. Some of this could be integrated into the science system but most should be on by default. 

3) Better mapping. I would LOVE to see SCANsat functionality in which the vessel actually collects data as it passes over the surface. It would be so much clearer to players why a polar orbit makes sense and creates a simple in-engine minigame thats actually about real spaceflight. Most important though would be to have those maps shown as direct overlays in map mode and flight mode so players could toggle through resource overlays, biome overlays, height and slope maps so they could pick smart landing zones and colony sites. It would also be helpful in flight mode as you're landing to have a graphical overlay showing you where the ground is flattest so you could more easily guide yourself down to safe LZ.

Edited February 11, 2023 by Pthigrivi

[GigFiz]

1 hour ago, Pthigrivi said:

1) An easy to use and understand Mission planner that lets players lay out their desired journey and provides them with dV totals and launch windows and lets them set alarms right there. Ideally the planner would be graphically simple and default to the next window, but offer drop down options to refine or alter launch dates using a porkchop plot. Even better would be the ability to save these mission plans to vessels or workspaces in the VAB so a player could look right there and see "Oh it looks like I'll need 430 dV to land on Dres but my lander only has 300... I'd better beef that up."

2) Better visual cues for precise landing much like what is found in Trajectories--visual targets and landing zone prediction, estimated difference between target and predicted landing point, trajectory factoring drag, perhaps even overheat prediction for reentry profiles. I'd also love to see some of the things Better Burn Time included like automatically showing time to impact/rendezvous and estimated burn time, estimated time to enter/exit an atmosphere. Some of this could be integrated into the science system but most should be on by default. 

3) Better mapping. I would LOVE to see SCANsat functionality in which the vessel actually collects data as it passes over the surface. It would be so much clearer to players why a polar orbit makes sense and creates a simple in-engine minigame thats actually about real spaceflight. Most important though would just to have those maps shown as direct overlays in map mode and flight mode so players could toggle through resource overlays, biome overlays, height and slope maps so they could pick smart landing zones and colony sites. It would also be helpful in flight mode as your landing to have an graphical overlay showing you where the ground was flattest so you could more easily guide yourself down to safe LZ.

Along these lines, something that you kind of allude to here, one thing that really needed to be done (and to be fair, seems to be something that has been in consideration, if not directly stated), was to take a comprehensive look at the mods in ksp1, take all the ones where the consensus is "you basically can't play without these", "you can play without these, but you shouldn't", and stuff like "this makes this part of the ui actually useful" or "this keeps you from having to tab out to go look up something on a website", and integrate their functionality in to the base game naturally (and hopefully improving/updating them where possible).

I (and I know Bej has said this before too) would love to have a mod folder that only got used for adding cool stuff or tweaking stuff in interesting ways, and have things be perfectly playable and fun with no mods at all 

[Scarecrow71]

10 hours ago, Bej Kerman said:

And is everyone into optimization? No, of course not. Some people aren't into doing rendezvouses all the time, others haven't even figured out how to dock. Point being, I see this argument all the time, as if a game like Kerbal Space Program doesn't need to be built to support reasonably sized rockets. It does.

Agreed.  I personally could use multiple lessons on optimization.  But do I go looking for them?  No, because I'd rather put a continent in orbit and play as opposed to spending a decade optimizing a ship.  

6 hours ago, Pthigrivi said:

An easy to use and understand Mission planner that lets players lay out their desired journey and provides them with dV totals and launch windows and lets them set alarms right there. Ideally the planner would be graphically simple and default to the next window, but offer drop down options to refine or alter launch dates using a porkchop plot. Even better would be the ability to save these mission plans to vessels or workspaces in the VAB so a player could look right there and see "Oh it looks like I'll need 430 dV to land on Dres but my lander only has 300... I'd better beef that up."

What I'd like is a dV calculator that tells me what the dV is of a particular section without it recalculating when I put stuff below it.  For example, if my lander has 400dv on it and then I put a decoupled on the bottom of it, don't recalculate the dV.  That decoupler isn't going to be there in flight when the lander detaches.

[Bej Kerman]

Just now, Scarecrow71 said:

Agreed.  I personally could use multiple lessons on optimization.  But do I go looking for them?  No, because I'd rather put a continent in orbit and play as opposed to spending a decade optimizing a ship.  

I mean, the egregious thing is that it doesn't have to be a continent. Just anything bigger or more exciting than a lander will do.

[JoeSchmuckatelli]

On 2/10/2023 at 7:49 PM, K^2 said:

Wheels is one of these things that everyone keeps getting wrong, even though it's not actually that complicated. Just a bit unintuitive in places. Get someone who knows what they're doing to implement your game's wheels. Intercept, if you're looking for somebody, I can do a contract. :p Seriously, don't just hand it to someone enthusiastic but clueless and tell them to figure it out.

Please go back to 2019 and see if you can get them to listen. 

Rover wheels were possessed in KSP. 

[Pthigrivi]

39 minutes ago, Scarecrow71 said:

What I'd like is a dV calculator that tells me what the dV is of a particular section without it recalculating when I put stuff below it.  For example, if my lander has 400dv on it and then I put a decoupled on the bottom of it, don't recalculate the dV.  That decoupler isn't going to be there in flight when the lander detaches.

Yeah I wonder if this is possible given their new subassembly system? 

[Bej Kerman]

54 minutes ago, Scarecrow71 said:

What I'd like is a dV calculator that tells me what the dV is of a particular section without it recalculating when I put stuff below it.  For example, if my lander has 400dv on it and then I put a decoupled on the bottom of it, don't recalculate the dV.  That decoupler isn't going to be there in flight when the lander detaches.

Is this a problem? Surely recalculating it every time anything changes only helps prevent issues where the game might have done a bad calculation and needs to re-evaluate the dV.