how big is the earth?
in General
Sounds like a dumb question. just google it, right? Bear with me...
More specifically, how big does Ingress think the earth is? And does it think the earth is a perfect sphere?
If I pick 2 points with lat/lon, how does Ingress calculate the distance? Still sound dumb? Bear with me...
I'm not looking for the EQUATIONS, i can get those from any maths textbook. I'm looking for the CONSTANTS.
Does Ingress assume a perfect sphere? of what RADIUS?
The equatorial radius of 6378km? The polar radius of 6357km? The global average of 6371km? The IAU nominal equatorial radius of 6371.1km? Some similar number with more bits of precision?
It starts to matter if you're trying to plan a 6882.28km link...
How big does INGRESS think the earth is?
Tagged:
6
Comments
I’m pretty sure it’s nearly a 1:1 scale. However, the Ingress App thinks that the world is flat, whereas the Intel Map understands there’s a curve.
it’s why “crossed link” situations like this bug report happen, as confirmed by Brian Rose https://community.ingress.com/en/discussion/comment/134186#Comment_134186
I wonder if @ofer2 can share some insights on what are the constants/calculations 👀
Might not be what was meant, but I've noticed some weird inconsistencies with the mu calculations, more than once, I've multi fielded and it will randomly drop in the amount I get when a field gets bigger, for example, I did one that was say 25k mu then go to the next portal about a kilometer away and it drops down to 22k .... how? Its still covering the same area as the first but has more houses added
That is to do with population density. It's calculated using S2 cells of a particular size. This causes the MU to change based on where the field is thrown from, not what is under it.
If it was based on what is under the field then fish units would be a real thing.
I should also add that my understanding is the population figures haven't been updated for a number of years, if ever.
So Ingress uses Googles library for calculating distances along the Earth. You can see the calculation it does here: https://github.com/google/s2-geometry-library-java/blob/master/src/com/google/common/geometry/S2LatLng.java#L220 . To answer your question more directly though, the library uses the Haversine formula to calculate the distance with a radius for the Earth which appears to be an average between the equatorial radius and the polar radius (6367000). As far as what the Ingress app shows, it was not built to show links as curved, but the server does treat them that way. I hope this answers your question.
@gazzas89 Yes, that is well known and has to do with the way that they are using S2 cell data with population density info for the calculation. Apparently crossing the boundary of a (mumble mumble) sized cell can make this happen occasionally. I can look up what mumble mumble is if you'd like.
Here's something else of note: Apparently there is (or was at one time) a different calculation for what's shown in your profile compared to what is shown for the MU in comm. I may well be the only person ever to have noticed this discrepancy... the comm message was 48,221 larger than what's on my profile, but that was on 1.6 trillion MU.
@ofer2 does Ingress actually calculate blocking links as though it it is on the curvature of the sphere?
I mean, if it didn't, all those curved links you see in polar regions would have a ton of crossed ones too.
Yes, it does.
We know from experience that it does. The only place it does things wrong is in rendering fields in the scanner. Well, they might have fixed that in Prime but in Redacted you could be under a field that didn't show up because the app didn't render curvature correctly.
so i would compute this one, for example, as 6881265.4m
do you agree?
https://intel.ingress.com/intel?ll=46.846465,-43.587059&z=4&pls=32.789015,-79.782613,59.700856,9.801937
Yes, 6881265.462785648 (so technically you would round up to .5 but w/e)
It was one of those questions I had many years ago, and tried to find some the math that would do it, found it last year, but the question if Ingress actually used it remained, or if a cheat was done. Because it has to be calculated fast, very, very fast.
Ofer2 thankfully answered that ponderous question.
Yeah, we do a lot of tricks to cheat, but those cheats are only used for performance and dont reduce accuracy. E.g. we segment the attempted link by s2 cells and then lookup all other links in those cells. This allows us to quickly and confidently remove most of the other links without having to check them. Its also why testing large links takes so much longer than small ones.
I am not a programmer, or a mathematician,however when I had thought about that problem, and thinking more and more about how slow the calculating could be, gave me a huge amount of appreciation of just how fast and optimised one of many parts of Ingress, we players don't give much thought into, and that we players just don't give credit, where credit is due to the developers in these areas of the game.
heheh, my actual was
6881265.44709588
but i think if we're quibbling over 0.015 meter, then i'm as close as i need to be.
also, in case someone searches for this in the future, IITC drawtools import json. adjust for your prefered start portal.
[{"type":"circle","latLng":{"lat":-30.000000,"lng":115.000000},"radius":6881280,"color":"#ff0000"}]
Now I’m curious what the longest possible link would be.
6881.28 km