EXECUTIVE SUMMARY:
- We build on prev work linking crop circles and UFO sightings.
- Several types of UFO's are strongly related to the appearance or
non-appearance of key features in UK crop circles. Some types of
UFO's represent only a small minority of all sightings but some
closely-associated types represent as many as 1/3 of all sightings
across N America.
- The predictive models found here are somewhat complicated and only a
computer would love. But the link between circles and UFO's is
highly highly unlikely to be just chance. We demonstrate one model
where the appearance of certain keywords and non-appearance of
others in crop circle reports predict around 1/3 of UFO sightings in
the following month.
- While the majority of circles may be hoaxes or "mistakes", like UFO
phenomena it seems a core of them are interesting and "mutually
validated" by certain types of UFO sightings in following weeks.


In a prev post we saw crop circles in the UK seem to contain
information about future UFO activity. Who is creating the more
interesting circles and whether the linkage with UFO activity is
intentional or incidental was not determined.

More number crunching has now revealed a vague association between
certain features as described for circles documented on <http://www.ukcropcircles.co.uk/>. And we can also establish an
ordering of which types of UFO activity are better predicted by those
crop circles. Roughly, we can decode what the crop circles are talking about.

At this point the patterns are no so simple as linking one circle
feature with a specific type of UFO activity. It seems the circles are
"talking about" more general things than a specific event at a
specific time and place. But the associations are nevertheless
statistically "strong".

We might use the information to decide which crop circles and UFO
sightings are more likely to be "real" unusual phenomena. Obviously,
like reported UFO activity crop circles appearing over the UK
countryside are mostly likely to be "mistakes" of one kind or
another. Those not created by hoaxers with boards and ropes may be
relatively in the minority of reports.

Breaking down UFO sightings (NUFORC) and crop circles into monthly
totals, adjusting for certain biases (e.g. in UFO sightings most
sightings are reported on Sats and the least on Mons -- we can
adjust the numbers to show what *should* be reported if people were
equally observant on every day of the week) we can build predictive
models for each UFO type based on the presence or absence of keywords
from crop circles seen in the immediately preceding month.

Putting that through a package that handles all the data reasonably
carefully (handling missing values appropriately, adjusting for serial correlations, robust statistical tests on each model and also on each
estimated coefficient in the models) we find the "top 10" models in
terms of predictive power are:

Model R2
dark 0.62260
-Light 0.60544
Fireball 0.59755
-orange 0.58111
Disk 0.56184
Fireball 0.54785
-Other 0.54650
-red 0.54426
-Cone 0.53796
-Egg 0.52568

The "models" are the features of the UFO sightings being
predicted. The R2 shows what% of the month-to-month variation
in UFO sightings are predicted from crop circles in the prev
month. All of the top 10 show the "majority" of UFO sightings (as
adjusted for various biases) are described by features in UK crop
circles. This might seem amazing given experts estimate only around
10% of UFO sightings in N America are anything "really interesting".
(I noticed an estimate from the French GEIPAN was even smaller, at
3%). But the explanation is the R2 applies only to those sightings
with the given "feature".

I.e. crop circles seem to predict 62% of UFO's next month that are
described as "dark". But "dark" UFO's represent only around 2% of all
reports. So it's only about 1% of all reports being predicted.

Interestingly, the 2nd-best group detected by this process are UFO's
described as other than "lights". Ufologists generally discount
reports of "lights in the sky" these days with all manner of
technology that can spoof various types. E.g. satellite trains, LED
kites, drones, drone swarms, etc.

But "not lights" is a fairly large group. About 50% of all
sightings. It seems crop circles predict 60% of non-lights UFO
sightings which represents 30% of all reports. Quite a big chunk!

It may be a subset of crop circles "validate" as many as 1/3 of all
American UFO sightings, far in excess of what experts previously
believed represented unusual phenomena.

Other groups may indicate even larger subsets. E.g. "orange" UFO's
represent around 18% of all sightings. It seems 58% of "not orange"
UFO's are predicted by crop circles features -- i.e. 48% of all
sightings. "red" UFO's are about 8% of sightings. About 54% of "not
red" UFO's are predicted by crop circle features -- around 50% of all sightings.

The models discovered by the s/w are, as described above, not all that
easy to understand. We can look at the "not red" model more closely
(given it describes 1/2 of sightings :).

The crop circle keywords from a given month that predict "not red"
UFO's next month in N Am are:

-central
-circular
-formation
ground
laid
large
like
many

I.e. the model shows the number of UK circles in a month that feature
keywords "ground", "laid", "large", "like", or "many", but NOT include
keywords "central", "circular", "formation" is strongly connected
with the number of UFO reports in the next month that do not contain
the keyword "red".

If we were expecting to decrypt the pictorial language of crop circles
we might be disappointed it is not one pictorial feature per UFO
keyword.

But it gets worse. Noone said the more reports with "ground" or the
fewer without "central" meant more "non red" UFO's next month. The
model allows the coefficient linking a crop circle keyword and UFO
sightings to be -ve as well as +ve.

The full details of the "not red" model are;

VARIABLE COEFFICIENT STAND. ERROR T - VALUE P - VALUE
----------------------------------------------------------------------
DATE 0.00008 0.00012 0.61284 0.54394 -central 0.07512 0.14152 0.53082 0.59889 -circular 0.43892 0.11970 3.66688 0.00081 -formation -0.39418 0.18711 -2.10670 0.04238 ground 0.96502 0.43684 2.20910 0.03381
laid 0.07426 0.06050 1.22748 0.22783
large 0.00415 0.07574 0.05475 0.95665
like -2.76230 1.28030 -2.15755 0.03791
many -0.67650 0.36273 -1.86501 0.07058
CONSTANT -9.48040 24.67817 -0.38416 0.70318

WEIGHTED SUM OF SQUARES = 3.48311
DEGREES OF FREEDOM = 35
SCALE ESTIMATE = 0.31546
COEFFICIENT OF DETERMINATION (R SQUARED) = 0.54426
THE F-VALUE = 4.644 (WITH 9 AND 35 DF) P - VALUE = 0.00044
THERE ARE 45 POINTS WITH NON-ZERO WEIGHT.
AVERAGE WEIGHT = 0.81818

The "F-VALUE" shows the model is statistically strong. There less than
1 chance in 1000 (the P-value) the association between this type of
crop circle in one month with UFO sightings in the next month could
just be due to lucky coincidence. We are "pretty sure" the 2 things are
linked. The R SQUARED says about 54% of month-to-month variations in
"not red" UFO sightings is predicted by the model (i.e. around 1/3 of
all sightings). And the "scale estimate" says the model predicts next
months total of not red UFO sightings within about +-30%.

From the "coefficients" we see the keyword "like" in a crop circle
description has the biggest effect. For each crop circle report in one
month that contains "like" there are almost 90% fewer not-red UFO's
next month. (A log transform is in effect and exp(-2.7) is around
.06). At the other end of the scale each crop circle in one month
with keyword "large" may be associated with 0.4% more not-red UFO
reports next month.

So it's a model only an AI would love. But suffice to say there is a
link between crop circles and UFO sightings that should a large chunk
of UFO sightings -- at least the ones at NUFORC -- are foreshadowed by
crop circles in the prev month.

As before, whether the information in crop circles is deliberately
planted or is some kind of accident cant be determined by these data.


--
"Online Haters" Are Psychopaths, Recent Study Finds.
The Debrief, 6 Apr 2021
Recently published research finds that "online haters" or people who
post hateful comments on online forums and social media demonstrate
high levels of personality disorder traits associated with psychopathy.

Ancient human migration into Europe revealed via genome analysis
The Guardian, 07 Apr 2021 18:30Z
Genetic sequencing dating back 45000 years shows intermixing with
Neanderthals more common than previously thought.

Oldest DNA from a Homo sapiens reveals surprisingly recent Neanderthal ancestry Nature.com, 07 Apr 2021 16:29Z
Scientists have sequenced the oldest Homo sapiens DNA on record, showing
that many of Europe's first humans had Neanderthals in their family trees.

UK infections drop about 60% amid vaccinations, lockdown
msnNOW, 08 Apr 2021 13:58Z
London (AP). The UK's COVID-19 vaccination program is beginning to break
the link between infection and serious illness or death, according to the latest ...

Masquerading, deadly snake discovered and named after shape-shifting Chinese goddess
Livescience.com, 08 Apr 2021 14:56Z
Scientists in China have discovered a new species of deadly snake and have named it after a shape-shifting snake goddess from Chinese mythology.

2021 Atlantic Hurricane Season Expected to Be More Active Than Normal,
Colorado State University Outlook Says
The Weather Channel, 08 Apr 2021 16:52Z
The 2021 Atlantic hurricane season is predicted to be more active than usual, according to an outlook released Thu by the Colorado State University ...

Using the human hand as a powerless infrared radiation source
Phys.org, 08 Apr 2021 13:51Z
A team of researchers at Shanghai Jiao Tong University, has found that the human hand can be used as a powerless infrared radiation (IR) source in multiple ...

IBM just launched the first developer certification for quantum computing ZDNet, 07 Apr 2021 10:22Z
IBM has unveiled its very first quantum developer certification, but what exactly does it suggest on a developer's CV?

CDC Says More Virulent British Strain Of Coronavirus Now Dominant In US
NPR, 07 Apr 2021 23:23Z
A more easily spread coronavirus variant first identified in England last
year has now become the dominant strain in the US, the head of the Centers
for Disease ...

Mysterious Plumes of Methane Gas Appear Over Bangladesh
Bloomberg, 08 Apr 2021 07:06Z
Satellites show large, frequent emissions over Dhaka-the latest discovery as scientists hunt for sources of the potent greenhouse gas.

[Yesterday is was just twins:]
And then there were three: Cyclone brewing off WA coast joined by two
tropical lows in rare weather phenomenon
ABC News, 08 Apr 2021 03:30Z
A cyclone brewing off the West Australian coast is joined by a second
tropical low, adding to what was already a rare weather event.

Real Life BB-8 Robot Will Explore Moon's Lava Tubes
The Debrief, 6 Apr 2021
Researchers developed a new drone, named DAEDALUS, explicitly
designed to explore and map the Moon's lava tubes.

Beijing now has more bnaires than any city
BBC News, 08 Apr 2021 02:50Z
The city has overtaken New York as it bounced back quickly from the pandemic, according to Forbes.

UPS reveals plan to buy 100s of helicopter-like electric aircraft to
buzz around cities delivering packages - take a look
Yahoo News, 07 Apr 2021 14:25Z
The big brown delivery trucks in the sky can carry up to 1400 pounds of
cargo and fly at speeds of up to 170 miles per hour.

[New physics!]
first results from Fermilab's Muon g-2 experiment strengthen evidence of new physics
Fermi National Accelerator Laboratory, 07 Apr 2021 15:45Z
The long-awaited first results from the Muon g-2 experiment at the US
Dept of Energy's Fermi National Accelerator Laboratory show fundamental ...

--- SoupGate-Win32 v1.05
* Origin: www.darkrealms.ca (1:229/2)