Talk:1172 Äneas

L5?
Aneas, based on its name, should be near Jupiter's L5 point, correct? I tried poking around the sources, but I'm not adept enough at orbital mechanics to tell if this is the case. 75.143.52.193 (talk) 09:32, 5 March 2016 (UTC)


 * Yes, (1172) Aneas is a L5 trojan. -- Kheider (talk) 16:52, 5 March 2016 (UTC)

"137km suggests too much knowledge, so reducing this to 1sf"
Don't be silly. The original sources specify a size out to 5sf (and an older version of the page would have you believe the NEOWISE data shows estimates out to 6sf, though I haven't been able to verify this in a satisfactory way other than it being mentioned in the JPL SBD with no explanation of how they found the additional sigfig from a dataset that only gives 5 when directly queried), with an error range of course. Three datapoints is of course a bit thin pickings, statistically speaking, but there's nothing wrong with taking the mean of those three then rounding it off to 3sf (or 2sf once converted to miles). It's a decent representation of the average of the available data, if we assume each of them to be as reliable as each other and thus equal worth, thus equal weighting in the average.

Why, if the scientists are each confident enough to publish a figure of that accuracy, can we not average them out to better than the nearest 10 miles (=16 kilometres)?

Over-rounding in this case essentially puts additional weight on the larger figures and less on the smaller, and as noted in the edit reason, leads to gross inaccuracy in the miles estimate (as it's only 1sf, vs 2sf for the kilometres - which if they were also 1sf would fall to an obviously very wrong 100km), to the point that it's actually suggesting a larger size (90 miles = 144km) than all but the largest of the three scientific measurements. Nearest integer unit in this case is the happy medium between suggesting excessive accuracy that doesn't exist (ie by reporting 2 or 3 decimal places), and rounding off the data so heavily that the figures are almost worthless.

We have to have some additional precision down at this level, even if it's a bit noisy (which is the truth of astronomical data, in the main), otherwise we end up with no real way to discriminate between the huge number of sub-150km-diameter bodies (and thus try to assign some kind of importance based on size, particularly when it comes to things that might end up coming near Earth) other than saying they're "probably bigger than 50km" or "probably smaller than 50km", even though there's enough bodies with decently correlated measurements down below the kilometre level.

If you really want, highlight the slightly wider than normal error range (ie 118 to 149km) by adding a plus-minus to it (in this case, $137$ would come close enough, or $136.5$ for additional precision), or parrot the Physical Characteristics section (which is a bit off-tone for the lede / abstract section) by giving the lowest and highest estimates. 146.199.0.203 (talk) 21:57, 12 March 2018 (UTC)