Is the radius of Kerbin _really_ 600,000m? RESOLVED: F3's pants are on fire

[seanth]

For whatever reason, the info screen (what you see when you press F3) is reporting distances 2x what they actually are. It's unclear as of this writing (2016.10.02) whether it always does that, or whether there is a minimum distance needed before it becomes incorrect.

Thx to everyone that commented and helped sanity check things.

TL;DR: I think there is an error in how KSP is reporting the size of Kerbin. Kerbin's radius can't be 600,000m because circumnavigations are longer than the expected circumference. Yet the geosync height for sats matches a planet with a 600,000m radius. 

For the past few months I have been refining a boat that will let me circumnavigate Kerbin without refueling. I based my initial designs on the information of the wiki that says the radius of Kerbin is 600,000m. That gives a circumference of 3,768,000m (2*3.14*600,000m). Based on the fuel efficiency of some early designs, I worked out a target fuel efficiency and the amount of fuel I would need to go 3,768,000m. 

The problem was, when I went that far, I was only about half way around the planet.

I didn't think too much about it and continued to refine my designs and increasing fuel until I finally managed the circumnavigation. The total distance in a test run I did was 7,866,453m. Again, I didn't think much about it. I was just pleased I had accomplished it. It wasn't until I started documenting things for a Elcano Challenge entry that I realized something didn't make sense.

What made it clear was that, for some graphics, I put a skin of Kerbin onto the globe in Google Earth as described here. The plan was to make isochronic maps in google earth to show my boat tests and refinements, but I needed to be able to convert distances reported when I would hit F3 in KSP to distances in Google Earth. What I would was the 2,500,000m on Earth would equal approximately 447,000m on Kerbin.

That would mean that Earth is 5.59x larger than Kerbin, not the 10.6x I had thought it was. If Kerbin's radius is 1,200,000, that would explain the Earth/Kerbin surface distance conversion and why my circumnavigation distance was 7,866,453m.

If I take the lat/long of the runway island (-1.517306, -71.965488) and the southern tip of the continent the KSC is on (-9.5581, -85.9790), you can plug them into online tools and get the distance on Earth between those points. That link reports it's 1789km (1,789,000m) between those points on Earth. If Kerbin's radius is 600,000m that means it is 10.6x smaller than Earth and the distance on Kerbin should be 168,773m from Runway Island to the tip. If Kerbin's radius is 1,200,000m, Kerbin would be 5.3x smaller than Earth, and the distance would be 337,547m. My notes say it was 384,410m during my most recent test run.

	If Kerbin's Radius is 600,000m	If Kerbin's Radius is 1,200,000m
Circumference	3,768,000m	7,536,000m*
Size relative to Earth	10.63x smaller	5.31x smaller**
KSC to southern tip (by water)	~168,773m	~337,547m***

* Matches what I have seen in my circumnavigation

**Matches what I have seen converting Google Earth distances to Kerbin distances

***Matches the distance I recorded in game.

I can't resolve what I'm seeing with what other people have reported in the past. For example, here's a circumnavigation done by a plane. It's total distance (3,876,325)matched what you'd expect with r=600,000m. And Geosync sats in game are at an altitude above sealevel that is correct if r=600,000m.

Can anyone else confirm this or tell me where I am being an idiot? I suppose a mod could be interfering, but the only mods I have are MechJeb and Hyperedit. Could there be something about traveling on water that is making the F3 info screen display the distance travelled incorrectly?

 

EDIT: In looking at F3 shots from people doing circumnavigations (this thread) there are differences in the distances reported.

• 25 June 2015: 4,189,588m (close to 3,768,000m for r=600,000m)
• 20 October 2015: 3,775,794 (close to 3,768,000m for r=600,000m)
• 19 February 2016: 8,308,562m (close to 7,536,000m for r=1,200,000m)

Could something have happened going from KSP 1.0.5 to 1.1?

Edited October 3, 2016 by seanth

[foamyesque]

The info screen is made of lies, basically. I never trust it to report accurately because I have seen it report significant discrepancies to fixed measures -- e.g. pieces left behind at KSC during a glide challenge.

[seanth]

36 minutes ago, foamyesque said:

The info screen is made of lies, basically. I never trust it to report accurately because I have seen it report significant discrepancies to fixed measures -- e.g. pieces left behind at KSC during a glide challenge.

That might be true, but here's another data point: When testing my boat, I would record speed, and mission clock time. I took 4 readings going from the KSC to the southern tip of that continent with the speed averaging 68m/s. It took me 1:48:40 (or 6520sec) to get from the KSC to the tip. 68m/s * 6520sec = 443,360m. That's further than the F3 reported 384,410m, but I know for a fact I wasn't going 68m/s the whole way. There were times when I would lose my hydrofoil would lose plane and crash down into the water, so my actual speed was slower.

I am getting more convinced this is some weird bug in KSP. In moving to unity 5, was there a mixup between radius and diameter?

Edited October 1, 2016 by seanth

[foamyesque]

I've done low-altitude flight circumnavigations (or partial ones, e.g. polar shots). My times and speeds on them are fully consistent with a 600km Kerbin.

The bug is in the info screen. For example, look at the reported distance travelled in that Feb 19 screenshot; it's roughly double the maximum speed * mission time value, which means one or the other is wrong.The mission clock is accurate, the speed-over-land is reasonably accurate, the distance number is pretty much always wrong.

 

Edited October 1, 2016 by foamyesque

[seanth]

Well, if horizons actually work in KSP, this can be answered by figuring out distance to horizon.

If the observer is 5m above sealevel:

• If Kerbin's radius is 600,000m they will be able to see 2.45km
• If Kerbin's radius is 1,200,000m they will be able to see 3.46km

(assuming my math is correct)

[Superfluous J]

1. Launch a ship.
2. Land it 180 degrees around Kerbin.
3. Put another ship on the launch pad.
4. Check the distance between them in map mode.

[SpannerMonkey(smce)]

42 minutes ago, seanth said:

If Kerbin's radius is 1,200,000m they will be able to see 3.46km

Hi, that actually backs up what i see all the time, I spend a considerable time at or just above sea level and reckon that around 4km will put a ship hull down when viewed from a similar vessel,  at 2.54 km the second vessel is clearly and completely visible, and apparently the BDAc radar agrees with me , prior to 3.5 km the vessel is clearly visible on radar, after 3.5km the vessel rapidly becomes invisible to radar as it drops below the horizon

[swjr-swis]

1 hour ago, foamyesque said:

The info screen is made of lies, basically. I never trust it to report accurately because I have seen it report significant discrepancies to fixed measures -- e.g. pieces left behind at KSC during a glide challenge.

This basically. Launch a vessel, immediately hit F3: 175m/s speeds reached, 10m covered/traveled, etc, and your vessel hasn't even had the time to let physics kick in yet. If you're in the habit of taking screenshots of top speeds, altitude, distances etc (for challenges or for testing) you'll constantly spot differences between what your in-flight screenshots prove you did and what the F3 screen of the same flight reports.

Treat that screen as purely indicative, no more.

[seanth]

5 hours ago, SpannerMonkey(smce) said:

Hi, that actually backs up what i see all the time, I spend a considerable time at or just above sea level and reckon that around 4km will put a ship hull down when viewed from a similar vessel,  at 2.54 km the second vessel is clearly and completely visible, and apparently the BDAc radar agrees with me , prior to 3.5 km the vessel is clearly visible on radar, after 3.5km the vessel rapidly becomes invisible to radar as it drops below the horizon

So that backs up the idea that Kerbin has a radius of 1,200,000m. Weird.

[Snark]

It's 600,000 km.  Lots of numbers back this up.

8 hours ago, 5thHorseman said:

1. Launch a ship.
2. Land it 180 degrees around Kerbin.
3. Put another ship on the launch pad.
4. Check the distance between them in map mode.

^ This was the immediate strategy that came to mind, and I was going to suggest it until I saw that 5thHorseman beat me to it. 

But there are other ways to check the math.  Examples:

• Measure it directly, as Horseman suggests above.
• Eyeball it.  Put a ship into a circular orbit at 600 km altitude.  Look at it in the map screen.  Is the size of the circular orbit about twice as far across as Kerbin itself?  Yes.
• Calculate it from orbital mechanics.  Put a ship into circular orbit at whatever altitude you like, up to, say, a thousand kilometers or so.  Make it a precise circle.  Note your orbital velocity.  Work out what the radius of the orbit would have to be in order to get that orbital velocity, then subtract your altitude in order to get the radius of the planet.  Result:  600 km.

 

[Rodhern]

I put a stock plane in the water and used as a boat. From the KSC heading east at the equator for 10 minutes at approximately 21 m/s, I traveled 1.21 degrees of longitude. Doing the experiment from around the zero meridian for 15 minutes eastbound at a little more than 21.5 m/s I traveled 1.86 degrees of longitude. Both results fit well with a radius of 600 km.

[seanth]

52 minutes ago, Rodhern said:

I put a stock plane in the water and used as a boat. From the KSC heading east at the equator for 10 minutes at approximately 21 m/s, I traveled 1.21 degrees of longitude. Doing the experiment from around the zero meridian for 15 minutes eastbound at a little more than 21.5 m/s I traveled 1.86 degrees of longitude. Both results fit well with a radius of 600 km.

I agree. Your measurements support a r of 600km.

I'm finding the whole thing confusing and frustrating. Sometimes the observations clearly say it's 600km. Othertimes they point to 1,200km. I guess for the time being I'll assume there is just something weird about my distance measurements. 

 

[numerobis]

The surface distance traveled seems not to factor out orbital velocity correctly, somehow. I have a save where there's a plane that splashed down after just over 7,000 km of flight. Flying slightly slower (200 m/s rather than a variable rate slowly accelerating from 200 to 250 m/s over a period of hours), when I cross that spot my plane has over 10,000 km of flight.

[Val]

 

23 hours ago, foamyesque said:

The info screen is made of lies, basically.

7 hours ago, numerobis said:

The surface distance traveled seems not to factor out orbital velocity correctly, somehow.

21 hours ago, swjr-swis said:

Launch a vessel, immediately hit F3: 175m/s speeds reached, 10m covered/traveled, etc, and your vessel hasn't even had the time to let physics kick in yet.

The measurements on the F3 screen includes the rotation of Kerbin.

Which makes it suited to measure how far a craft has traveled, while it is in actual orbit, but less than useful for surface measurements.

I think, the F3 screen is made that way, because having it switch between surface and orbit relative measurements, would introduce it's own problems and inaccuracies.

You could make a rule to switch above a certain altitude or when leaving atmosphere, but that'd introduce large inaccuracies for suborbital jumps that just cross the line for only a short time.

You could make the F3 screen show both surface and orbit, but what happens if you land on another body. Which surface should it show measurements relative to. Should it now show measurements relative to both surfaces? What if we land on 10 bodies.

Sorry, for rambling.

[swjr-swis]

2 hours ago, Val said:

Sorry, for rambling.

Rambling's fine , as long as you understand I am not complaining about the F3 screen; it is what it is.

I was just providing a quick reality check showing that there is no point to expect the numbers on that screen to be 100% accurate or anything but purely indicative.

[SpannerMonkey(smce)]

Something that may give a hint of some smoke and mirrors type behavior and perhaps some explanation of the visual anomalies regarding horizons  was something i found in the 1.2 1553 change log namely "  Extend the visual horizon a little. " which seems to point to some system in which the visuals aren't tied exactly to the  planetary dimensions

[Kerbart]

31 minutes ago, SpannerMonkey(smce) said:

 something i found in the 1.2 1553 change log namely "  Extend the visual horizon a little. " which seems to point to some system in which the visuals aren't tied exactly to the  planetary dimensions

Instead og getting the tin foil hats on, I think it is simply meant as the distance beyond which objects are not rendered, for performance reasons.

[SpannerMonkey(smce)]

46 minutes ago, Kerbart said:

Instead og getting the tin foil hats on, I think it is simply meant as the distance beyond which objects are not rendered, for performance reasons.

No tin foil hat here,  just an observation.

[Kerbart]

36 minutes ago, SpannerMonkey(smce) said:

No tin foil hat here,  just an observation.

I don't think claiming, and I quote,  smoke and mirrors type behavior is an observation. If so, show some evidence of that. Extraordinary claims require extraordinary evidence, claiming that Kerbin is a different size than listed and that smoke and mirrors are used to cover that up... It passes the smell test for me regarding a "C" theory.

[seanth]

10 minutes ago, Kerbart said:

Extraordinary claims require extraordinary evidence, claiming that Kerbin is a different size than listed and that smoke and mirrors are used to cover that up... It passes the smell test for me regarding a "C" theory.

I hope you don't think I'm trying to present bad data. I know the claim that Kerbin might be a different size is pretty out there, but the data I presented in OP is what I have gathered thus far supporting Kerbin being a different size. I acknowledge there are other observations that support the accepted size, and I can't reconcile why some people are seeing data going one way and why other people are seeing data supporting something else. Maybe there is some weird bug in F3 distance calcs depending on whether you travel E-->W vs W-->E? Maybe deviating to the N or S impacts it? I don't know. All I know is I have numerous observations saying Kerbin is larger than I thought it was. Having said that, I have orbital data saying Kerbin is the normal size. Sooo...this thread.

I have some ideas about looking at time of sunrise at the KSC and then time of sunrise very far away (but still on the equator) along with the horizon distance experiments. This week is very busy for me, but I will try and compile all the observations I have. I really need to come up with a way of obtaining distance measurements without relying on F3, though (thus the time of sunrise at various points on the surface idea)

[foamyesque]

1 hour ago, seanth said:

 I really need to come up with a way of obtaining distance measurements without relying on F3, though (thus the time of sunrise at various points on the surface idea)

Target mode and hyperedit

[Red Iron Crown]

13 hours ago, Val said:

The measurements on the F3 screen includes the rotation of Kerbin.

Pretty sure this is the root of the confusion here. Kerbin's radius is well established through several methods, not least of which is the programmatical definition of the planet itself. Start with the assumption that the F3 screen is in the inertial reference frame and not the rotational one and all these "inconsistent" observations start to make sense.

Another method to confirm this is to stop Kerbin's rotation using Hyperedit or similar tools, suddenly F3 becomes accurate for surface travel.

[Leszek]

Your orbital indicator with Kerbal Engineer makes it clear you have a 600 KM kerbin since you can never have a PE of less then -600.  A number of other factors make it clear that Kerbin has a 600 KM diameter.  For example I regularly put satellites in 1K orbits.  1.5K orbits and so forth.  This would just be about one radius out from the planet on a 1 200 KM Kerbin but they are obviously much farther out than that.

[Kerbart]

Of course there are inaccuracies with screen measurements, but based on these 2000x2000km orbits I'd say Kerbin has a diameter of 1200km (I leave it to the student to figure out that  17.3×71 is indeed close to 1200)

[Gaarst]

Yeah, if Kerbin was not the right size, I think we'd have noticed by now.

13 minutes ago, Kerbart said:

Of course there are inaccuracies with screen measurements, but based on these 2000x2000km orbits I'd say Kerbin has a diameter of 1200km (I leave it to the student to figure out that  17.3×71 is indeed close to 1200)

I was about to go there and do an angular size measurement to prove Kerbin has a radius of 600km, but you did it for me