reply to post by yankeerose
Hi Yankeerose and thank you very much for posting that prediction info. I hope you don't mind if I make some comments on it.
For those who may have arrived on this thread only recently (and welcome to you!), a little explanation could be helpful. There are many places on the
web that offer earthquake predictions. Some do more-or-less what we do here (but it seems generally without so much follow-up), while others are
rather absurd, posting huge numbers of predictions for quakes that cover just about all seismically active regions of the world on an almost weekly
basis, often with little or no specificity as to time frame or magnitude. That is not prediction at all, any more than predicting rain in around
50-plus known "rainy" locations around the world each week. Of course you will get some hits but they are meaningless. To be reasonably valid,
predictions need to state locations, give a reasonable time frame or at least only hold valid for a reasonable time (say a week or two at the most,
usually), and either state a range of magnitude or automatically exclude quakes below a certain magnitude.
The general approach here is that for an area with quite intense seismic activity, like Indonesia for example, we do not claim a "hit" for a quake
that is (by that region's standards) quite mild. Indonesia gets many sub-5 mag quakes, and even those in the low 5 range occur a few times every week
on average. So any predictions for such regions take this into account. On the other hand, places like the US West Coast see very few quakes over a
mag 5, so it is reasonable to claim hits for quakes below that range. Here, though, we tend to be more specific as to location. The smaller the
expected quake, the more specific its location. That's about the way it usually works. Even so, while we might report very small quakes in the 1 or 2
mag range, they aren't claimed as hits unless they occur in very unusual places and were predicted to occur.
Another point is that this thread was not established by Kattraxx and Damntheptb as a "prophet of doom" kind of thing. It was set up to consider and
evaluate various methods of prediction, exchange ideas and keep tabs on results. We will not find the best methods or even learn very much by only
making predictions for massive, "killer" quakes or volcanic activity. They are thankfully quite rare and the number of such predictive posts, if
valid, would not run more than a couple of pages. Instead, it is the ongoing and developing process of trying to improve prediction methods for a wide
range of seismic events that is the key. This also means that while we take note of "hits", we also consider what we can learn from "misses".
Very few other sites I have visited seem to have much interest in discussing or reviewing their "misses". They just ignore them. This is
unscientific and does not make for reasonable and ongoing research. It also suggests a certain level of denial on the part of those who run such
sites.
Now moving on to Miriam's prediction, it can be evaluated with no difficulty. The time frame is quite specific and narrow, she defines a reasonable
geographical region, and she gives an expected minimum magnitude.
What's the chances of her being right?
Statistically, there are globally around a dozen mag 7-plus quakes every year (give or take a couple). Her time frame is only around three days, which
is 1/10 of a month. If I've worked this out right, then, the odds of there being a mag 7-plus
anywhere in the world within her stated period
are about 10%. In other words, if you were to predict that a mag 7-plus quake will occur anywhere in the world within a stated three-day period -- but
with no specific location stated -- you'd get a "hit" about one time out of ten. Those are not great odds and if you were to do it purely on a
random basis then you'd look pretty woeful.
But that's not the end of it. The odds of such a quake being within her defined geographical region are clearly far less; there are at least ten
locations I can think of around the world where on the basis of historical data, a mag 7 would not be totally out of the ordinary or at least is a
physical possibility. Some of those regions are far more seismically active than the stated region, and hence more likely to have a major quake. Just
working on rough figures, though, I'd say the odds of accurately predicting a 7-plus quake in that stated region within that time frame are not much
better than 1%. Yep, if you predict a mag 7-plus quake for a specific region in such a narrow time frame, you'll probably be right about once in a
hundred times. If you choose a very active region you might get that down to one in fifty.
My own feeling tends towards her being very close with this prediction; I have a level of concern about the stated region and those who've followed
this thread for any length of time will understand what I mean when I use the word "concern". While I do not, however, see indications of a 7-plus
by August 3 in the stated region, something in the high-six range may be on the cards yet it could occur slightly outside the stated time frame. I'd
go for something by the end of August 5. (Local time zone, just to be clear.) I'd guess the odds of my being right would be around 16% or so, as mag
6-range quakes are about ten times more likely than mag 7 ones and I've extended the time period by two days. In any case I'm only giving my opinion
and it will not trouble me at all if I'm wrong. It's all data.
Some readers might see little difference between a high-six (say a 6.5 to 6.8) and a mag 7-plus, but in terms of the amount of energy released the
difference is quite enough to be of significance.
Mike.