Originally Posted by
BayAreaPilot
I mentioned my suspicion earlier in the thread, but I am now pretty convinced that the discrepancies are the result of using a great-circle formula that assumes Earth is a sphere instead of the more accurate Vincenty's formula that assumes Earth is an oblate spheroid.
If you want to test for yourself, use
http://www.gcmap.com/ to get the result using Vincenty's formula. Then paste the lat/long of the airports into the boxes at
http://www.movable-type.co.uk/scripts/latlong.html, which uses calculations that assume a spherical Earth.
Now whether this was intentional or simply an uninformed programmer using the wrong formula is up for debate. As they say, never attribute to malice that which is adequately explained by stupidity. But it certainly is possible UA did a study and found using the less accurate formula would lead to fewer miles being awarded. And given that nearly every route changed, I would hope the changes would have never made it through QA without UA management signing off on them.
Checking it against
my earlier post, it seems like this is the most likely explanation. I get a result for LAX-NRT using the sphere formula of 8754 km, or 5,439.48 miles. Very close to the 5,437 generated by the new formula, so maybe a couple of calculations that were rounded off in the process.
In fact, the website BayAreaPilot provided states:
[In fact, the earth is very slightly ellipsoidal; using a spherical model gives errors typically up to 0.3% – see notes for further details].
This means that errors from assuming spherical geometry might be up to 0.55% crossing the equator, though generally below 0.3%, depending on latitude and direction of travel.
I must have fallen asleep in geometry when the differences between spheres and oblate spheroids were discussed, but I'd guess that the routes that changed positively are generally north-south in bearing and closer to the poles would gain the most from the switch?
My NRT-HKG route gained 11 miles, despite it being a much shorter route than the closer-to-the-Equator HKG-SIN route, which only gained 3.
Though I have to disagree that this was intentional -- differences of a few miles on routes would probably the least of the integration team's worries when reviewing the functionality of the combined system. Even if they noticed the minor changes, they might chalk it up to rounding, an inconsistency in the two airlines' databases, or some other reason, and move on, choosing to spend their time focusing on the most-important issues.
In hindsight, it can generate millions of savings for United (since, if my hypothesis is right regarding direction of travel, there are fewer passenger seat-miles in generally north-south routes than there are in east-west routes), but I doubt that UA management (1) noticed this and (2) signed off on it.