swansont

Physics Girl, Veritasium, MinutePhysics, SmarterEveryDay, The Slow Mo Guys, It's Okay To Be Smart. First time poster, long time lurker.

B. John Jones is suspended a week for a rule 1 violation on top of repeated thread hijacking and soapboxing (an apparent inability to post without bringing religion into it)

Velocity_Boy is suspended a week for repeated thread hijacking and abusive behaviour.

bart2 has been banned as a soockpuppet of Bart.

Mikemikev has been banned for his insults, his refusal to follow moderator direction, soapboxing and persistent use of fallacies in his arguments.

Again, supply and demand. The people that show up here discussing an invention are usually very coy about details because they're afraid someone will steal their idea. Often that's overblown because they have a fundamental flaw in their understanding of the science. Another issue would be whether the discussion counts as prior art for patent purposes. I have no idea what the answer to that is. The result is that there isn't much call for it. You can e.g. discuss mechanical or electronic details in physics or engineering.

Prophet12A has been banned as a sockpuppet of Prophet12.

I had a chance to peek at some of the supporting data here. The premise is that we can treat everything as random for the null hypothesis. However... The first quake+riot is described as "Second wave of violence 3 weeks after earlier Toxteth incident" Doesn't sound like it's random. It's correlated with an earlier event. The next two instances happen on the same day. Both are "Violence between striking miners and police", so the riots are correlated with each other. Even in the most generous treatment of this analysis, they should not be counted as separate incidents. Further down: two instances of "Copycat disorder during weak economic conditions" More correlated events.

Oh, *That's* what you're calculating. What are the size of the error bars on this prediction of 9 riots in this period of time? I mean, you can do this with coin flips. You expect as many heads as tails, but in any 10-toss stretch, you can easily get 6 of one and 4 of the other. Or 7 and 3, somewhat less often, and 8-2 less often than that, etc. you only get 5H and 5T about a quarter of the time. 6H happens about 20% of the time. 7H is more than 10%. So how do you conclude that the odds are less than 1% for your riots distribution?

Yes, my mistake, I had that backwards. 9 riots in a two week period. That's even worse, since you only have 64 of them. You claim 21 riots per earthquake actually happened, before and after. That only requires 3 earthquakes. Something is not adding up. The bins I'm referring to is just any random two week period. Surely you analyzed a null hypothesis that there's no correlation at all. The expectation of a riot in any period is less than 0.1 and yet you have 16 of them before an earthquake. Again, something doesn't add up.

I don't believe that to be true. The sampling error is the square root of the sample size if it's white noise. Seeing as your data points are around 2 deviations from the average (9±3, so 2 deviations is 15 and 3), so being substantially less than 1% doesn't seem like a reasonable conclusion. And that excludes any bias which would give you a systematic shift. Are there any other 14-day windows with such large and small counts that aren't associated with an earthquake? ——— How do you get 9 earthquake per two-week period, anyway? 151 quakes in 33 years. There are 26 two-week spans per year, so there are 858 bins in the data set. How do you get 9 quakes per period from that? How do you get 21 riots per earthquake (16 before, 5 after) with 151 quakes and only 64 riots?

As a resident of DC I can observe that benign quakes have an outsized effect, since buildings in quiet areas are generally not built with quakes in mind. We had a 5.8 quake here a few years ago. Lots of damage. We got laughed at by the southern California crowd, because for them that quake is not unusual. (Turnabout being fair play, they freaked out when the temperature dropped to freezing a couple of years later) How close are the earthquakes to the riots? Klaynos mentioned small data sets, which looks to be a big issue. Are your results statistically significant? If so, at what level?

If things were randomly distributed. But the aftermath of an earthquake might occupy peoples' attention, making them less likely to engage in acts of civil unrest.

overtone has been banned. Since racism has been added to previous transgressions (hijacking, abusive behavior, soapboxing, failure to take directions from mods when asked for evidence), we decided that's enough.

There was a report on a locked thread in the Lounge, citing the "Discuss life, work, school, anything!" description, implying that the anything literally means that anything can be discussed. That's ridiculous. The rules still apply. You can't use the Lounge to post things that we would crack down on elsewhere for rules violations. It's not a consequences-free zone. Further, and I really don't think this should have to be pointed out, the implicit meaning is "anything you wouldn't be talking about in the categories we have set up", which is a broader application of the desire for posts to be on-topic and relevant to a discussion. Posting in the Lounge rather than in an appropriate sub-forum is lazy and a bit rude (especially if one does it habitually)* If you want to be free to post anything you like, start up a blog or tumblr somewhere. It's pretty easy to do. *edit to add: posting in speculations for similar reasons is equally lazy.

shmengie has been suspended 7 days for repeated thread hijacking

! Moderator Note New thread merged with existing discussion

Robbitybob1's ability to create content has been removed for 2 days, so that he may contemplate how to stay on-topic and refrain from hijacking threads. Edit: this has been upgraded to a full suspension

redkiss has been banned as a sockpuppet of rusek and various other incarnations of Maciej Marosz (some of which were spam-banned and thus erased)

From a recent report: Closed threads aren't necessarily due to rules violations; they can simply run their course. If the OP is not amenable to modifying their claim in the light of evidence and we keep running over the same ground, we are going to close the thread. There's no point to continued discourse. Warning about an infraction is a courtesy, of sorts, especially if one has had similar discussions closed before, as in this case. In that case it makes pleading ignorance farcical.

Atlas Hyperion has been banned as a sockpuppet of SigmaR.I.F.T_CKF (and of VehGalTal)

We get about 1 inquiry a year about why e.g. there is no economics forum. So it's a question of supply and demand.

In response to a request, MonDie has been granted a six-month sabbatical.

Frank Martin DiMeglio has been suspended for two weeks for spamming us with the same topic, after repeatedly being told not to re-introduce the subject.

puppypower has been suspended 7 days for repeated thread hijacking and soapboxing.