This is another of the W.O.O.D. series of semi-regular
Weekly Occasional Open Discussions.
(i.e. if I forget and skip one, no big)
Immediate prior one here:
and remains open for threads running there
(at least until the ‘several month’ auto-close of comments on stale threads).
Canonical list of old ones here:
So use “Tips” for “Oooh, look at the interesting ponder thing!”
and “W.O.O.D” for “Did you see what just happened?! What did you think about it?”
What’s Going On?
I’m back from a flash-marathon coast-to-coast-to-coast dash of about 5700 miles in about 2 weeks… Oh Joy – NOT! Don’t ask why. There were a couple of things with time dependency that I needed to do in Florida plus 2 back here just after them, one yesterday… Yes, it would have been more effective to fly, had I time to book 3 months in advance and if rental cars were not such a PITA these days… The world has made air travel pretty crappy when it is more convenient to drive coast to coast than it is to fly (deal with security lines, TSA “Your PAPERS Please!!”, getting there 3 hours early, having your fight canceled, spending the night on a bench in the Denver Airport due to wind in the Rockies, eating crap from ‘whatever is open’ at 3 AM, spending an hour getting a rental car to find out you have signed your life away, etc. etc. and getting to pay $800 for a $200 flight because you didn’t book it 4 weeks in advance on a maybe…) But everything is now taken care of for another year.
Yesterday was the California DMV. I renewed my California driver license. They seemed to know I had an out of State license (asked “Have you had an out of State license?” to which I answered truthfully “yes”) but didn’t say my California license had been canceled or anything so maybe it was still valid. In any case, it was a standard ‘renewal’ process plus written test. I failed the Motorcycle “written” (on a computer now) the first time as they have nutty ideas of what part of the lane you “ought” to ride in at various times. My rule is “ride in the best least oily part of the lane that lets you avoid accidents and tells cars it is MY lane and I don’t share well” and that varies dramatically with the texture of the road and what is going on around you. They have “fixed ideas”. Often in conflict with the MC Class mandated for my Son (I sat in on it with him) where they properly said to avoid the grease / oil stripe in the middle of the lane most of the time.) Then there were the nutty “all of them are right” choices. Like “What matters to safety?” with choices of what amounted to “situational awareness” said badly, your protective gear, or knowing your motorcycle. Well Situational Awareness matters right up until the unavoidable happens, then it’s “do I know this bike enough to avoid that jerk?”, then it’s protective gear time. They ALL are most important just at different times. So, OK, count down from 120 so the 2 minute timer to retake expires and then passed it on the second go with their stupid answers.
Now I get to wait 2 weeks for the license to arrive in the mail (no, they can’t actually make one on site, unlike Florida…) then I’ll check in with the online Florida site and see if this caused the Florida one to be cancelled. I’d originally just wanted to find out IF my California license was still valid (in which case I’d renew) or had it been cancelled via getting the Florida one (in which case I’d have just ‘let it go’ and kept the Florida one as it doesn’t expire for 4 more years anyway and I’m going to be permanently there soon enough). But there was no way I could find on line to find out “current status” and on arrival at the DMV I got tracked into the renewal process so just “road it out”. It used to be that Florida would just give you a license as a snowbird for the 6 mos you were there, and you didn’t end up with the constant flip-flop from one State to the other. Now, post “Real ID” requirements, there’s this situation where 10 days (or maybe 2 weeks) after arrival at either end you are REQUIRED by law to change licenses (which now lots of folks just ignore…) so in theory 2 x each year the Snowbirds get to do the “DMV Thing” all over again. For me, it would have been 4 times in one particular year. Just stupid. So maybe I’ll still have both, or maybe in a few more months I’ll be back at the Florida DMV…
I’ve not kept up on the status of BREXIT. IIRC it’s supposed to pick up the process again in a week or so. The Yellow Vests were partly placated, but went ahead and protested anyway. I need to find out their status too.
Trump had a meeting about The Wall – under military rules. I suspect so anyone who leaked could be put in the slammer under UCMJ rules ;-)
I still like the idea of just having the Army practice “road building and barrier erection” all along the border. There’s a new crop of recruits who get to learn all that every few months anyway (they presently build / tear down over and over again on bases…) so just assign the Military to “protect the border”.
Stock market has been on a very nice up run from a down spike, but is Not Yet Confirmed as a reversal to the upside.
Price is above PSAR (those red dots) so a trade-in situation from that last down spike. But only now is it reaching the SMA Simple Moving Average stack. This is the time of testing. In a down trend (SMA stack inverted with fastest on the bottom) the presumption is it returns to the SMA stack and then has another down leg. This holds until such time as prices managed to “punch through” the SMA stack, return to it from above and bounce off to the upside. Only then does the “trend bias” swap from Bear Market Rules to Bull Market Rules.
Looking at volume, it is FALLING as prices rise. Not good for the Bull Case… There was a good ‘trade-in’ opportunity back where MACD went to “blue on top” and where DMI had the Red cross the Black ADX line; HOWEVER we are still MACD below zero and DMI has ‘red over blue’ so still Bear Market Rules. Trades end when prices return to the SMA stack – then you decide is it another Bear Run (so trade short) or do we get a confirmed reversal?
So far, it’s just looking like a garden variety up-spike in a down market trend. (Eventually these will always reverse, but you simply can not know the number of “down legs” before the reversal becomes real. You can GUESS or you can WAIT FOR CONFIRMATION but you can not accurately PREDICT.)
With that, I’m back at my workstation and can once again post in comfort ;-)
FWIW, while transiting the country, I was pondering the best Database Structure for bringing all the various temperature datasets into one easy to compare place. I think I may have a decent idea on it, but it will take some manual work. The basic problem is that the prior structures used by the climate folks use Country as part of the ordering / key and countries change over time; plus they use “WMO” number but that has a couple of variations over time ( original was shorter, then they added a ‘version’ for when there were a few thermometers at a place, now they have ‘homogenized away’ the version number). I was thinking that an “ascension” value to denote just which / when NCDC / Hadley / whoever data set it was, then a WMO+Version key to the particular Thermometer or Homogenized would likely do it. So XXXXXyyZ where XXXXX is the WMO number (that seems to not change much) and yy is the Version for those sets that have it, and where Z is either V or H for ‘with version’ or ‘homogenized’ and with other flags possible as needed.
I THINK that would let me load all temperatures from all the data sets (with things like ‘date of ascension’, country code and continent and all as dependent values rather than key portions) for all versions. Then comparing the change of values over time and over datasets would be fairly simple to code. Maybe ;-) More on this “someday” as I putter on it.
I also put some think time on the problem of “shaping their ends”… Part of the “mystery sauce” IMHO is likely to be that they start life for a given instantiation of a thermometer in a cold period, then end it in a hot period. Now that cyclical sine wave becomes an “uptrend” and is ‘averaged in’ forever to the overall (bogus) warming ‘trend’. So how to detect that, and fix (sorry “adjust”…) it? Well, I had two thoughts. One was to fit a trend line to a given thermometer data set, then prune the lead in and exit back to the trend line. It shortens the history but reduces (doesn’t fully eliminate though) the entry and end biasing. Measuring how much of that was needed would say something about how many sets start low and end high vs start high and end low. Similarly, the “trend before” vs “trend after” would also say something about trend lines fitted to cyclical data with chosen starting point and end point…
The other thought was just to plot a scatter diagram for each thermometer for each “month” over years. So, for example, have ALL December data for Casper Wyoming plotted as temp divergence from trend up/down and year right/left. Then look at it. Is it more or less random, or do the early years diverge below trend and the later years above? Is there a cycle in the middle and the ends set the trend? Repeat for the 1200 major thermometers (so 14400 graphs) and maybe then the 6000 ish total (so about 72000 graphs) then come up with some automated way to evaluate them ;-)
Thoughts? Ideas? Likely a few months before I “go there”… and start writing some kind of code. It would be nice if there were some prior art on detecting bogus trend lines in cyclical statistics via end effects; but I’m not in touch with the field enough to know what search terms to use for my search-foo on this one.