View Full Version : Nonagon Revealed

07-02-2010, 07:59 PM
To start things off, I know that some people have said that Nonagons actually roll better than circles. I found this really hard to believe at first so me and my friend nitewalker11 decided to do a few tests to prove/disprove this theory. And just to calrify things, a Nonagon is a nine sided shape, and a circle in LBP is actually a 20 sided shape. So, here is a complete guide to Nonagons.

How to Make a Nonagon
Coming soon, with pictures!

Our Preliminary Findings
After we made our Nonagon, we started to fiddle around with. We immediately noticed that the Nonagon rolled smoothly; possibly even better then LBP's circle of the same size. Also, when we tried to make the Nonagon parallel to the ground, but no matter what we tried, the Nonagon would not sit perfectly on one of its nine sides. This led us to believe that LBP rendered a Nonagon as a perfect circle, rather then a nine sided shape.

Nitewalker11's Test
Hey all, nitewalker11 here.
For my test, I used a 42 small grid high, 127 small grid long ramp. (No, these weren't planned numbers, they were randomly gridded out) I added a backstop so that I could line the figures up perfectly in the same spot. At the bottom of the ramp, there was a magnetic switch with a radius of 25 small grid, attached to an emitter. The emitter emits small, cardboard squares into a pit. I put a magnetic trigger in the center of both figures. I allowed both the circle and the nonagon to roll down the ramp, through the magnetic switches, and counted the number of squares emitted. The circle had 22, and the 21, proving that the nonagon rolls faster than the circle.

Marauder47's Test
For my test, I used the same ramp as nitewalker11(a 42 block high, 127 blocks long, with a 133.7647 block hypotenuse). I then lined the circle up against the backstop and then pushed it against the ramp while in pause mode. I then had a .1 emitter which would emit one block every .1 seconds into a box. I then created a finish line which had a proximity swich and then put it on its side and changed the angle to 180 degrees so that it was a straight line. I then attached the proximity switch to a permanent switch that would turn off the emitter when the ball crossed the finish line. I then placed another proximity siwtch on the center of the ball that would activate the proximity switch at the end of the course. I then unpaused the level so that the ball would roll down the ramp and perfrom the experiment. I then repated these steps for the nonagon. In my experiment, I proved that the Nonagon was faster. Over the time it took the circle to go over the course, 88 blocks were emitted. Over the time it took for the Nonagon to go over the course, 84 blocks were emitted. This means that the Nonagon was 4.8% more efficetn then the circle.

In case you didn't notice, Me and nitewalker's results were proportionally similar,therefore proving that nonagons consistantly roll nearly 5% faster than circles, and with less than half of the vertices
If your looking for smoothly rolling vehicles, use a nonagon. And, if you want to cut back on thermo, nonagons will do any job that circles could, with less vertices.

Pictures of Making a Nonagon





Pictures of Our Experiment




07-02-2010, 08:21 PM
Shameless bump.

07-03-2010, 01:27 AM
Breaking news!

As it turns out, Nonagons are recognized by the game engine as "Perfect Circles." While doing some more in depth research, we saw that perfect Nonagons actually float and are never able to touch the ground. But, if you severly corner edit it or add another side, the object falls right to the ground. This leads me to speculate that there is an invisible "force field" around the Nonagon that actually allows it to roll better and makes it a "Perfect Circle." We proved this by placing an object into the force field whithout attaching it. But, the object was suspened inside the force field and made no visible impact on the Nonagon itself. Now the reason why I put Perfect Circle in quotes is that at this point, we are beginning to think that the game engine recognizes perfect Nonagons as a shape with many sides that are unknown to us at this time. As a side note to this experiment you can put 1449 Nonagons into a level without it overheating. But, only 1339 circles can fit into a level. This would lead us to believe that the Nonagon has more then 9 verticies, but less then 20. But, the thermometer allows the same amount of triangles, squares, and octagons into a level so, it has different rules when regarding vertice caculation in a regular polygon. So, we cannot prove this theory. Further testing will be done. As of now we still have no comments, so we would enjoy some constructive critcism or new ideas.

07-03-2010, 01:37 AM
Kind of hard to do anything until you tell us how you made it... ;)

Also, your test of smoothness by rolling it down a ramp proves nothing. Larger objects fall faster, as it is simulated gravity. This just means that your nonagon was bigger than your 20-sided circle, assuming that both were treated as spherical.


07-03-2010, 01:47 AM
Plug: I have more in-depth tutorials about protracting here (http://www.lbpcentral.com/forums/showthread.php?20656-Complex-Shapes-Tutorials-The-Visual-Guide&p=361691#post361691) if anyone's interested.

I think you're right; nonagons do behave like circles. You're not really the first to discover it though. I first found out in Lite_Sleeper's level called "LBP Physics Oddity: Polygon Race."

Also, the LBP circle also has an invisible forcefield around it that makes it behave like a circle. Make a circle huge and run on the edges if you want to test it.

07-03-2010, 01:53 AM
Kind of hard to do anything until you tell us how you made it... ;)

Also, your test of smoothness by rolling it down a ramp proves nothing. Larger objects fall faster, as it is simulated gravity. This just means that your nonagon was bigger than your 20-sided circle, assuming that both were treated as spherical.


As far as I'm concerned, they both weigh as close to exactly the same as we could possibly get them, we used a low tension spring as a scale. we changed their sizes until they both weighed the same and rolled them then.

But if you make a nonagon using the sprung bolt, and making each vertice (vertici?) exactly 40 degrees apart, then you'll see that the nonagon develops a force-field type of invisible force, similar to some Mm objects. It is impossible make one side flush with a flat surface because of the field.


By just a very minute, maybe a 50th of a small grid on our spring scale, the circle actually weighed more then the Nonagon. So therefore, we were actually testing the Nonagon agaionst a heavier object, and it still performed quite a bit better. Proving, that the Nonagon is slightly more efficent thermo wise, and speed wise.

07-11-2010, 09:57 PM
I made one myself using the sprung protractor technique, and I am quite convinced that the physics of the game treat it just like a circle. Whether it actually rolls with less resistance I think is arguable (even if the nonagon has less rolling resistance, it is not an impressive difference), but I do like that it seems to reduce thermo. :D

EDIT: You might want to mention that when you create a nonagon, you have to use the corner edit tool to make the full shape out of the 9 rectangles you see in the picture. Took me a little while to figure that one out.