Page 1 of 1

Experimental Track: "Invert"

PostPosted: 11 Feb 2019, 12:41
by tempAnon093
I want to start off with a clear disclaimer. This track is an experimental proof-of-concept. It is not intended to be in STK main. It violates many important STK style guidelines. It is not polished or pretty, and the outside lane of the track is glitchy. (middle and inside lanes are pretty safe though!)

Video: (it took a lot of takes)
Blender project:

So, let me introduce you to this wonderful track:
The start.
Oh, that's disappointing...
Yes, I am new to Blender and have no experience making materials or textures or trees or any scenery.

But don't worry, it gets better.


A loop!
Yes, with the magic of magnets you can drive upside down without fear of falling. I haven't seen a track yet using magnets for anything past 90° banking, so hopefully a better modeller than me can use this as inspiration.
Those who have made tracks before may notice the checkline at the top of the loop. We wouldn't want you taking a shortcut through the middle, would we? (although I did manage to throw a bowling ball through the middle, hitting out the kart in second place.)

Before you get to the loop, there's some other exciting track parts.


That's an inline twist! Notice that the support structure hangs over the top, because you drive along the bottom. Maybe someone more skilled can make a Mobius strip track using a half twist instead?
I can't make cutscenes yet, so here's gameplay screenshots of it: (They would look better if I had time to do proper decoration)

Look at that strange minimap!


So, about that small white arrow sign after the twist.
After the twist, the Bezier curve I used to create the track was twisted at almost 360° and for the sake of the driveline I assume it has to go back the other way to return to 0°. But two of the same twist might be redundant. So I made this... thing.

I don't know what it's called.

I don't think it even has a name. Let's call it an inverse banked U-turn ribbon, or something. Whatever it is, it's fun.
The reason for the arrow is so you have a bit of warning before the track vanishes from under you. When driving the track quickly in reverse, you can fly into the air into orbit around this curve and somehow land on the track, if you are lucky.

? ? ?

Anyway, after that is the loop, as seen before. And to end all this excitement, there is a simple tunnel.

Are those Nitro cans on the roof? Yes.
"But how can we reach them if they are on the roof?", a player may ask. That's alright. I made a tutorial.

Yeah, you just drive up the wall. Just make sure to get back down before the tunnel finishes.

Can you even drive on this thing?
Yes, but it is not perfect. Chances are this is because of my lack of understanding of how to make smooth tilted road in Blender (Cocoa Temple did it fine).
I tried to make the road very high-poly and smoothed the Bezier curve I made it with, then checked the STK option to smooth normals. As a result, the inside and middle of the entire track is fine to drive on. However, the outer-most part of the twisting corners is rough and will cause collisions and may even knock you backwards or into the air off the track. You shouldn't be driving there by choice though, that's the slowest lane of the track!
After a couple of laps to get used to the track, you can complete the circuit at high speed without track collisions or falling off, on any difficulty.

Can AI drive on this thing?
They're great at it! I've seen no issues with AI yet. UPDATE: on some difficulties a couple of them may drive too fast on the strange inverted-banked curve and fly off the track.
I didn't make the roof of the tunnel an alternate driveline, so they won't drive up those walls.

Any other known issues?
The track (and some other small objects) is 2D and appears one-sided due to how the graphics work (I guess), so the opposite half of the loop may be invisible while driving towards it, along with some other twists and turns in the distance.
If a properly made track had these same road features, this likely wouldn't be an issue, but I didn't know an easy way to fix it.

Feel free to try the track out, it is rather fun I think. I have also uploaded the blender project.

Re: Experimental Track: "Invert"

PostPosted: 12 Feb 2019, 23:33
by QwertyChouskie
Pretty cool! Reminds me of something XGhost posted a little while back on IRC. Maybe kill the grass and make the track in space instead. I'd love to have the blend posted, this could make a great addon with a bit of work.

Re: Experimental Track: "Invert"

PostPosted: 13 Feb 2019, 00:39
by tempAnon093
QwertyChouskie {l Wrote}:Pretty cool! Reminds me of something XGhost posted a little while back on IRC.

Would you happen to have a link? I am interested in seeing that track! If the link is lost, a simple description of the track would also be nice.

QwertyChouskie {l Wrote}:Maybe kill the grass and make the track in space instead.

That is a good idea. I did not think about theme or scenery at all before starting (and I suppose the default grassy theme is a habit left over from Rollercoaster Tycoon).
A space theme would really suit the modified-gravity aspect of the track.

QwertyChouskie {l Wrote}:I'd love to have the blend posted, this could make a great addon with a bit of work.

I will try and do that within 24 hours. However if someone would be able to remake/smoothen the track twists to make it less glitchy, that would be excellent (the video shows what sometimes happens at high speeds or on the outside of the track during twists).
I have extremely little experience with Blender, so the track may not be as easy to modify as a regular Blender project.

Thanks for the compliments! I mostly made the track because I haven't seen many take advantage of the special track properties like magnets and cannons, so any add-ons you know that already include these would certainly interest me.

Re: Experimental Track: "Invert"

PostPosted: 13 Feb 2019, 01:03
by QwertyChouskie
The track itself was never posted AFAIK, only this recording: (link from

Maybe I'll PM him to try to get a copy of the blend.

Re: Experimental Track: "Invert"

PostPosted: 13 Feb 2019, 01:07
by QwertyChouskie
OK, just PMed.

Re: Experimental Track: "Invert"

PostPosted: 13 Feb 2019, 12:12
by tempAnon093
Here is the Blender project:
The big dark plane in the sky was a hacky way to create a track-wide shadow, since I didn't want to learn proper lighting in the short amount of time I had.

This track is absolutely free for anyone to do anything with, although attribution is appreciated (as "GumballForAPenny"). STK's textures have their own licenses.

QwertyChouskie {l Wrote}:The track itself was never posted AFAIK, only this recording: (link from

Maybe I'll PM him to try to get a copy of the blend.

Thanks, that track certainly looks fascinating (I do agree with the IRC chat that it wasn't made to look pretty!) It would be good if you get a reply from XGhost.

Re: Experimental Track: "Invert"

PostPosted: 16 Feb 2019, 17:16
by dlewan
I've just started playing STK and must say I'm very impressed with the design. Specifically, I love the add-ons. (Although I doubt I'll ever have the artistic talent to create a decent one myself.)

Well, this post has answered a question that's been on my mind: Would it be possible to create a non-orientable track? Apparently, the answer is "yes".

Does anyone know of any such tracks? What about tracks based on famous manifolds? Boy's Surface? The Klein bottle? Calabi-Yau? How about minimal Riemann surfaces?

Re: Experimental Track: "Invert"

PostPosted: 17 Feb 2019, 15:46
by acme_pjz
dlewan {l Wrote}:Calabi-Yau?

No, Calabi-Yau manifold is 6-dimensional as a real manifold, and STK is not a 6D kart racing game :cool:

Re: Experimental Track: "Invert"

PostPosted: 18 Feb 2019, 01:43
by XGhost
@QwertyChouskie Got your pm, still lurking around. Thanks for the heads-up anyway!

Hi there, tempAnon093

First off, I do not have the track from my recording anymore. I'll try to find a blend on my workstation in the next few days, but I think it is gone. Sorry about that.
This track was one of many experiments for testing possible and interesting track layouts or track features. Like you, I see much unused potential in the current set of tracks and dove deep down into track analysis and track design for kart racing games. I really want to make a long post someday in the future, sharing my findings and trying out some stuff by creating a track from scratch. But up til now, I didn't have the time or investment that would be necessary for such a project. I'll list some things I think are helpful and maybe I can give you some insights into my research. Some things are not exclusively track design related, rather general game design practices, but they're useful anyway.

First of all, I would like to distinguish between the terms 'track layout' and 'track feature'. As I see it, the track layout is the general course of a track. The thing that defines how hard a player needs to steer to get a curve, how long someone can build up speed before the next turn or how well shortcuts or alternative sections are integrated. It is all about balancing player handicap with challenge. The track layout is the thing the player can least reason about most of the time. For example, if a player often loses control in a specific curve, she/he might say something like "this turn is too tight, make it wider". And while this may be correct, most of the time the actual problem is not the curve itself, but maybe its banking, something blocking the players sight (noticing curve too late) or the layout of the outgoing section before the curve (unable to reach the ideal racing line before entering the curve).
A track feature on the other hand is something special that makes driving the track unique and interesting. Think of movable obstacles, ramps, holes... even cannons or magnetic sections. These things sometimes directly affect the track layout and change how a track is driven. In some cases a 'not-so-good' track layout can be improved by adding track features or maybe a track really just lives from a well-done feature. However track features are tricky. A poorly designed feature as well as too many features can become punishing very quickly. Take the golfball obstacle in the minigolf course for example: The obstacle is fast, rolling over the whole track with no visual announcement when it spawns, right after a right turn (camera catches the obstacle only when it is too late) and if hit, the player will be thrown into the air. Although a cool feature on paper, it feels punishing.
Most of the time an idea for a new track originates from an idea for a track feature.

Now, as a rule of thumb I would say, if your track layout itself feels unsatisfying to play, fixing it with features will likely not work out. I did an exercise I think every future track designer should do once in a while:
The basic task is to create a track that is just a basic circle. One main driveline, one checkline, export, play. Obviously this will be a pretty boring track. Now in a second step, make this circle interesting. Alter the angles of the curve, add some banking, experiment with the width/height and draw an ideal racing line. In the end you should have a basic and very primitive track layout, but one that drives well. Doing that helps one understand how to tweak a layout to the better, maybe you'll find some best practices for different sections.
Man, there would be so much to cover just talking about layout. Balancing, feedback, playfeel and so many more awesome things. However I want to refer back to what you actually are working on ;) .

So then, magnetic sections...
Your 'invert' demo track is clearly a nice prototype, testing different possibilities of road bending and trying out various features. You may have noticed the drawbacks of magnetic sections in your tests, I will elaborate on what I have experienced and how I think they should be used in track designs.
So let's start by look at how a magnetic road in STK actually works. Basically it's pretty simple: the normal of the surface will be the (inverted) gravitation for the player and (and this is important) the up vector for the camera. This therefore means, that as long as you have no visual reference to the surrounding world, the player may never notice the effect of the magnetic section. This is good, but sometimes also bad. One of my tests with magnetic sections was a track, where I built a small road leading through a village. The catch was, that the whole track was modelled upside-down with magnetic roads. Ingame the player can not see what is actually going on. Then there was a ramp in the track pushing the player away from the road, leaving them to gravity's will. Suddenly the karts flew up to the air (meaning the ground) and revealing the actual nature of the track.
Now if you want to use magnetic roads for the effect (screw elements, loopings, etc.) you want to avoid this most of the time. Your goal is to get the player amazed about the crazy twist the road is going to make. So you need references that subconsciously guide the player on what they are currently doing. In the background, one should be able to tell, that they are currently driving stunts. Make it clear where up and down is. Work with the sight of the players camera, let them see in which direction the path will go, before they enter it. In your track, the screw element is not bad. When driving on it, the player constantly changes orientation and because the camera lags behind, this gives you the tilting effect, making it clear what is happening.
That's me driving your track (no skidding):
Now you see, when I exit the screw section at [16s], the row bends downwards, making it impossible for me to see what's coming next. Left curve, right curve? I don't know, so I will reduce my speed (also, I barely can control my kart on full speed here). Yes, there is an arrow indicating a left turn, but is it really a turn? Look at the track layout (more about the layout later), you can see I just hit the left arrow on my keyboard twice for a small correction. If I'd have turned, I had fell.
Next there is ground control (to major Tom). You surely noticed that magnetic roads are a pain to balance in terms of speed and the centrifugal force affecting the cars. The thing is, a magnetic road has always physical slope of 0, leading to maximum possible kart speed no matter how the layout is laid out. When bending roads, the gravity will be changing direction for every face of the mesh and this can press the karts onto or away from the path. This makes the thing also harder to control. I would suggest to let the stunt elements take up more space (longer screws and bending roads) and reduce the karts speed by adding preparatory curves before these sections (track layout).

I think the hardest part about magnetic sections is to get the track layout right. Even though I consider those effect-loaded sections as features, they heavily change the players perception on how to interpret the road. If you look at your track, you'll see the underlying layout is basically just oval. You can see in my run how I barely needed to steer while driving the screw or the looping. This is because the layout is basically straight there. It feels like a turn but it isn't. And that's the tricky part... I never managed to get a magnetic road to be really interesting to drive, because it is so much harder to setup a layout that works in 3-axis space. Look at the 90 degree turn in Cocoa. 90 degree -> this is not a curve anymore, because of manipulated gravity this is a straight line after the initial bend. For your screw element the same -> it is a screw, but all I do is driving straight. But it feels like I would need to steer. The first runs on Cocoa I fell into the water many times when taking the 90 degree section. This is because I'm getting ready for a curve that turns out to be none. I think this is the key for successful magnetic sections. Getting the track layout to work well.

I attached a screenshot from my track recording and added some comments which hopefully can explain my deliberations. There is a reason the road is bending that way.
  • Upwards, facing to the player, to let them see what is happening and how the karts in front will drive to the tunnel.
  • Make a curve (left) to reduce speed before the overhead intro to the tunnel
  • Switch direction (right curve before tunnel) while heading upwards to keep layout interesting

The tunnel then is quite funny. It is intended that the player loses all references for up and down. At this point you can basically layout the tunnel in any direction without the player ever noticing what's happening and where they are. I even textured the grid slightly screwed to make the players that follow the quads drive a full circle in the tunnel. What's not present in this track is the proper exit of the tunnel -> you need to redirect the players early in order to get them coming out all at the some position. But this is a topic for another day.

Hopefully I could give you some insights. Always open for discussions and questions.

Re: Experimental Track: "Invert"

PostPosted: 19 Feb 2019, 00:47
by QwertyChouskie
Excellent post! MTres, maybe link to this post on the track making wiki? It'd be a shame to have such valuable info just buried in a forum somewhere.

Re: Experimental Track: "Invert"

PostPosted: 19 Feb 2019, 23:25
by tempAnon093
Thank you for the detailed post, XGhost! As QwertyChouskie said, this is an excellent post that could provide a lot of tips to track designers. If you have the time, it would be very helpful to create a thread on the main forum page aimed at general track development to share the knowledge you have gained from your studies. There are a lot of good tips there about track layout and user experience especially which would apply to every track designer.

As you said, I do feel there is much unused potential in the current tracks, such as magnetic track features such as screw, loop and tunnel elements as well as dynamic hazards (I have written a short post about them, suggesting that a well-placed, timed hazard such as falling rocks or a moving train could be used to help challenge the faster expert players but not the slower players). My track is a demo or prototype for trying a couple of novel track features and it may be clear that I put no planning at all into track layout (when banking and incline is removed, it is a very simple oval). This track could definitely benefit a lot from a properly planned layout, for example making a player slow down before entering the parts that are dangerous at high speeds and having a moderately challenging driveline.

One point you mentioned is how in your own track there are lots of visual references to suggest what will happen next, such as seeing where the track is leading to next so the driver can prepare. My demo does not have much indication at all, the road has no pattern or railings on the side to show depth and direction and the driving mesh is one-sided so that half the loop is hidden from the driver. Proper decoration would help fix this, and modifying or removing elements such as the curve at [16s] would also help.
There is a small amount of environment (grass terrain, sky, supports) that can help the player notice when they are upside-down, but this could be made more obvious and more enjoyable with better scenery.
A note about the curve at [16s], I realised when making that element how driver-unfriendly it was. I agree that it is not a good track element to put in most tracks (especially due to how tight it is and how it was placed after a high-speed stretch of track). The track pulls down and away from the kart without any visual indication and the warning sign I made is misleading as the magnets reduce the amount of turn needed to drive on the track. I personally find it fun and interesting but would not recommend it in a proper track which should be clear and easy to drive.

You make an interesting point about how if a curve is banked enough (eg. the 90 degree curve in Cocoa Temple), it is just straight track or driving upwards for the player. It is probably possible to make a banked oval track so that the driver can simply not turn the entire lap, a very boring track I would say. It makes me think about creating banked or inverted elements where the driver is not fully guided by the magnets, such as creating a corkscrew, cobra roll or similar element but adjusting the track's normal so that the magnets only affect pitch, rather than pitch, roll and yaw. This would leave the driver responsible for turning left or right (or even forwards and backwards ;) ) I notice that you had that type of inversion in your track leading into the tunnel, where you have to drive to the right to stay on track.
Another example of having a track feature be more engaging is having a tunnel like your own track did as well, where freedom is given to the driver to take their own driveline. This is why I placed Nitro cans on the roof of the simple tunnel I had; a player can choose to drive at high speed through the middle (maybe they have boost, or are a beginner player) while another player may try to drive up onto the roof to collect a Nitro can reward. While this wasn't planned ahead by me, it is a track feature which also acts as a track layout influence, introducing an optional challenge as an alternate route. I also think tunnels and pipes would be easier to place into many tracks without it seeming strange and out of place.

Again, thank you for your excellent reply. It would be good to have many of the points you mention be used as STK official recommendations to help people design tracks. Important details such as forcing a player reduce their speed before a tight turn or using banking to reduce the perceived tightness of a corner are all good tips to give to people. It's easy to see decorations as simply eye-candy, so the notes about forewarning are also good to share.

Re: Experimental Track: "Invert"

PostPosted: 21 Feb 2019, 01:38
by eltomito
This reminded me of a track idea I had which was driving on contrails (those white vapor lines jet planes leave behind them) from one continent to another. One continent would be landmarked by The WIlber Tower and the other one by The Statue of Pinguerty. Sorry for the bad artwork but you'll get the idea. Yeah, you get back to the old continent by jumping from one aircraft carrier to another.