00:14 <egg|cell|egg> Bofh: meow

00:20 <bofh> egg|cell|egg: no idée

00:21 <egg|cell|egg> Bofh: :-\

00:21 <egg|cell|egg> But how do I uncertainty

00:23 <bofh> I'm trying to read some papers on this topic, hence my lack of response earlier :P

06:31 <egg|cell|egg> Bofh: meow

08:30 <egg|zzz|egg> bofh: hmmm, if I take the sequence m: k -> average of q_1 .. q_k, then the autocorrelation-based estimator for the standard deviation of the average of m looks like a decent estimator of the standard deviation of the average of *q* ???

08:31 <egg|zzz|egg> bofh: is there some sanity to this,

09:27 <egg|zzz|egg> hmmm http://adsbit.harvard.edu//full/1996AJ....111..541F/0000546.000.html

09:28 <egg|zzz|egg> (author whom i found via https://tamino.wordpress.com/2011/07/22/how-not-to-analyze-tide-gauge-data/ citing his book on light curves)

09:28 <egg|zzz|egg> not sure how but i ended up back at astronomy, after going through econometrics and tides

10:51 <egg|zzz|egg> bofh: hm perhaps the estimator being shitty is because of a poorly-chosen cutoff of the sum of the autocorrelations

10:52 <egg|zzz|egg> bofh: otoh if I put it at n/4 I occasionally get negative variances, which, um.

10:52 <UmbralRaptop> imaginary standard deviation? o_O

11:00 <egg|zzz|egg> UmbralRaptop: i have no idea what i'm doing, etc.

11:11 <UmbralRaptop> hrm. GR in 3 hours.

11:22 <UmbralRaptop> ;8ball will I get stabbed by a light conesnail?

11:22 <kmath> UmbralRaptop: Nope

13:18 <UmbralRaptop> https://twitter.com/sjs917/status/1087994883334389760

13:18 <kmath> <sjs917> Graham (@doccosmos) demonstrates the field of view of ZTF, which is quite a bit larger than other comparable survey… https://t.co/aSxjrqAwC6

13:18 <UmbralRaptop> … did LSST get delayed?

14:44 <Ellied> UmbralRaptor: I think this camera only shoots DNG and JPG

15:11 <UmbralRaptor> I mean, I was mostly being silly. >_>;;

16:00 <UmbralRaptop> Ellied.uptime++

16:29 <Ellied> UmbralRaptor: indeed

16:29 <egg|work|egg> UmbralRaptor: how was GR

16:41 <UmbralRaptop> egg|work|egg: relatively okay?

16:41 <UmbralRaptop> There's some questionable stuff going on with the homework, though.

16:44 <UmbralRaptop> Also, an edition of one textbook was changed to foil my using sci-hub.

16:46 <UmbralRaptop> (Additional problems)

17:34 <UmbralRaptor> George Mason deploys his robot army (2019, colorized)

18:00 <e_14159> egg|zzz|egg: I have fairly strange question: Do you think ICBMs account for the gravitational influence of other bodies in their guidance software?

18:36 <SnoopJeDi> e_14159, I'm no ovum, but I can't imagine they do, unless you count correction on final approach. If you're on Twitter, I bet @NuclearAnthro or the other weapons history people know the answer from declassified weapon specifications.

18:37 <SnoopJeDi> and even then, the reentry vehicle is *mostly* concerned with atmospheric drag and countermeasures, I think

18:40 <e_14159> That's what I had assumed

18:40 <SnoopJeDi> the trajectories are AFAIK suborbital and extremely fast, so there's not a lot of time for really finicky subtle effects (which I gather most such things are) to make much influence

18:41 <SnoopJeDi> UmbralRaptor, yea it'd be slightly shocking if the reference geoid wasn't part of the theory underlying the guidance package

18:42 <SnoopJeDi> but the defense industry also gave us https://twitter.com/pomeranian99/status/858856994438094848 so ¯\_(ツ)_/¯

18:42 <kmath> <✔pomeranian99> Memory leaks on missiles don't matter, so long as the missile explodes before too much leaks. A 1995 memo:… https://t.co/jZSQcbmXRK

18:43 <UmbralRaptor> More CEP == more megatons

18:46 <SnoopJeDi> yea, philosophically you just swallow guidance/intercept problems into SIOP/whatever-its-called-now

19:02 <UmbralRaptor> Anyway, a 9 megaton airburst looks to have a 14.6 km overpressure (5 psi) radius

19:13 <egg|zzz|egg> bofh: meow

19:16 <egg|zzz|egg> bofh: uncertaintea

19:19 <e_14159> There are targets that require a slight bit more precision.

19:20 <e_14159> egg|zzz|egg: Your opinion as resident n-body expert?

20:13 <egg|zzz|egg> e_14159: not sure how third-body effects would be relevant; you certainly want oblateness; maybe higher degree and order geopotential (although since you're interested in only a fraction of an orbit even that might not matter)

20:14 <e_14159> egg|zzz|egg: That certainly makes sense, thanks.

20:15 <egg|zzz|egg> meow

20:16 <e_14159> It might be a trout.

20:18 <egg|zzz|egg> e_14159: can you into statistics

20:19 <e_14159> egg|zzz|egg: It depends, but in all likelihood not. What's the problem?

20:21 <egg|zzz|egg> I want to estimate the error of the mean of some sample of a time series that ought to be stationary; obviously that's not independent, so I can use some autocorrelation-based estimator of the error of the mean, but it's not really марковesque, it's probably periodic (but I don't care about modeling the periodic behaviour, just estimating the error of the mean)

20:22 <egg|zzz|egg> e_14159: something like https://i.imgur.com/SAgBNRi.png

20:23 <e_14159> Unfortunately, I can't into that

20:23 <e_14159> But are you sure it's not марковesque?

20:25 <e_14159> At least from your example you might no... ah, no, the underlying long-term change wouldn't be modelled. Nevermind that then; am stupid

20:28 <egg|zzz|egg> e_14159: I mean in the eggsample there's a fairly obvious periodic signal

20:28 <egg|zzz|egg> but I just care about the mean

20:29 <egg|zzz|egg> which is easy to take, but then I want to have an error on it

20:30 <e_14159> The mean is over all samples, right?

20:35 <SnoopJeDi> egg|zzz|egg, hmm, maybe there's a dependence inasmuch as modern guidance systems will probably use (but not rely on?) GPS data, which itself has these kind of nuanced correction factors embedded

20:35 <SnoopJeDi> cc e_14159 ^

20:36 <egg|zzz|egg> e_14159: yeah

20:37 <e_14159> egg|zzz|egg: In that case, shouldn't it be independent of a reordering of points and therefore not matter?

20:37 <e_14159> SnoopJeDi: That might be. I'll still consider it an argument won ;-)

20:38 <egg|zzz|egg> e_14159: ? I want an estimate of the error of the sample average as an estimator of the mean of the actual (infinite) time series

20:38 <egg|zzz|egg> whitequark: oh they even have a twitter account https://twitter.com/Betty_Bourbaki

20:39 <whitequark> nice

20:43 <egg|zzz|egg> whitequark: which has just been followed by bofh and fibmod apparently :-p

20:50 <e_14159> egg|zzz|egg: Ah, then I misunderstood you.

20:50 <e_14159> More tomorrow, I have to zzz now

20:53 <egg|zzz|egg> have a good zzz

21:32 <bofh> https://twitter.com/bofh453/status/1088184715163721733

21:32 <kmath> <bofh453> Mon voisin Totoro. ⏎ Carrières Souterraines de Paris, Décembre 2018. https://t.co/32TZ9gJUHm

23:33 <UmbralRaptor> AAAAAAAAAAAAA https://twitter.com/astronamir/status/1088195483846303744

23:33 <kmath> <astronamir> PSA to colleagues: @NASA NSPIRES will be "unavailable" after Jan 31 because of the federal shutdown – there's an (e… https://t.co/XHzeYNAEYs

23:59 <UmbralRaptor> "Because Neugebauer and Leighton were physicists, they had no idea what astronomical results to expect. “I went to the astronomers and asked how many stars will we see? They said 75,” Neugebauer remembers. “That was the biggest number I got.”8 Neugebauer and Leighton weren’t the only ones to be surprised when their Two-Micron Sky Survey, which they published in 1969, gathered light from 20,000 stars."