Can earthquakes be predicted?

Total Pageviews

Search This Blog

Wednesday, November 09, 2011

Results of Year 2011 Earthquake prediction ,till October 2011

hi
Here are the actaul quakes (7+)  as against predicted dates for the year 2011 ,till October 2011
1) January
  Prediction date -------------actual dates and quake
   4th------------------2nd at 2020Hrs UTC----7.2
2)April
     8th-----------------7th at 1432 Hrs--7.1
3)July
   9th------------------6th at 1903 Hrs---7.6
                                   10th---------------7.0
4)August
 22nd----------------20th at 1819 Hrs-----7.1

5)September
   5th-------------------3rd at 2255 Hrs UTC---7.0
   16th-----------------15th -----7.3
6)October
   22/23rd-------------21st--------7.4
                                   23rd--------7.2
  Thus there are 9 Hits of 7+ quakes
 If you do not consider little inaccurcy of 2/3 Hrs ,there are 5 dot hits (+ or - one day window and 7+)


There are 19  quakes of 7+ during January to October 2011(incluning clusters)
The average number of 7+ quakes per year are 15
thus odds are 15/120----120 because ,total days 360/ 3 day window period
                       ie 1/8   ie one in eight or 0.125
 Now 9 hits out of 19 is 9/19  0.473
 Even if we consider 5 hit out of 19 it comes to
              5/19 ie 0.263
Both way it is above random date selection or average occurances(even after taking in to account cluster of quakes)

Amit

20 comments:

Anonymous said...

Respected Amit Sir

Congrats. It's a good percentage. We know that scientists demand 99% accuracy which is unreasonable because nature is still a mystery ( especially quakes , volcanoes and weather ) . This is because many forces and causes apply, some of it is gravity , tectonic stress and pressure at fault lines etc.

We also know that in year 2012 there will be possibly great quakes from Apr. to June 2012 on Earth around full moon days ( + - 2 days ) and new moon days ( + - 2 days ) . Locations can not be predicted yet only some LATITUDES ranges may be predicted which are from 8 to 38 deg. North of equator and 4 to 34 deg. south of equator . LONGITUDES ranges can not be predicted.

best regards

Santosh Phadnis
______________

Roger Hunter said...

Amit;

You have only 4 correct predictions out of 20 for the first 10 months of 2011.

You don't have a Jan 4 prediction on file.

Multiple quakes don't count. We're scoring predictions, not quakes, unless you predict the number of quakes in the window.

Your score is not significantly better than chance.

Roger

AMIT said...

Hi Roger
pl see my earlier post of 7+ quakes for the year 2011, where in date 4th was mentioned below the list ,because the window period was appoarching at the time of poating ( poated on 2nd January

Roger
the date 4th January 2011 is also there on your archieve blog( foot note below the list of dates)

http://quakepredictions.weebly.com/amit-dave.html

Amit

hi
As stated earlier,please note the dates for probable major earthquakes (7+). The window period will be + or - one day (UTC). Figures in brackets shows 7.5+ quake on that date only.
1) January 2011------19,25
2)February-----------19,21
3)March-------------- 20(7.5+)
4)April----------------3,17(7.5+)
5) May---------------16,29
6)June---------------9.12,15
7)July----------------1,9,31
8)August------------6(7.5+),13,30(7.5+)
9)September---------3(7.5+),12,16,28
10)October----------27(7.5+)
11) November-------24
12)December--------7,10,19,25(7.5+)

Total 28 dates hence 84 window dates ie 23.3 % of the year
There is one chance of 7+ on 4th January 2011.However as we are very close to window openiing that was not posted
Depending on cycle quakes ,occurring 7 days in advance, few dates may be modified (well in advance of window opening)
Please watch dates
Amit

AMIT said...

Roger
The original question is
Whether the dates predicted by me are better than random dates given by any body ?
or
Any better than average occurance of 7+ quakes?

Amit

Roger Hunter said...

Amit;

The odds on a 6.7 or larger quake in a 3 day window are 20.9%

If I allow your Jan 4 prediction yo would have 5 hits in 21 tries which is 23% while 4 hits in 20 tries is 20%.

So you are not significantly better than chance.

Roger

AMIT said...

Roger
So it is better than odds.
We have not deducted clusters due to 11th
march Japan quake.
Amit

Roger Hunter said...

Amit;

You are not enough better to matter.

If you toss a coin 10 times and get 6 heads are you better than chance?

Yes, by one head.

Does it matter?

No.

Next time you might only get 4.

Roger

AMIT said...

Roger
It has proved that for major quakeso7+ the prediction is well above average
By that logic ,what if I get 9out of 10 or say 10 out of ten.It coulcd be also by chance.where do we stop?
In our engineering field we test samples from lot and declare whole lote as approved or even rejected.Are there any chances that 5% samples are good by chance and there could be defective material in rest of the 95% ?
Amit

Roger Hunter said...

Amit;

Yes, it's possible.

I suggest you look up the z-binomial test for significance. You can run it with different numbers and get a feel for how many are needed to get a given level of significance.

VassarStats has an app for that which I use.

Roger

AMIT said...

Roger
If my predictipns are by chance anybody can get the same result ie above odds for 7 plus quakes
You have been evaluating quake predictions for so many years and for so many predictors like Petra,shown, etc.Have cpme accross any one crossing the averagefor major quakes of seven and above?
Amit

Roger Hunter said...

Amit;

Yes.

Petra Challus came very close to passing the test, so it's not impossible.

Your problem is that the method you're using just isn't valid.

Roger

Unknown said...

Hi Amit,

I'm certain Mr. Hunter is gravely mistaken given he failed to factor in catalog errors, so I no doubt passed his so-called test and he's afraid to admit it for fear his colleagues will eat him alive.

However, despite his comments the results were sufficient to bring about an offer to literally go out on fault lines and chase quakes as a TV series, yet Don and I had to decline given California is in a quake drought. But who knows, one day when things pick up an audience might be able to see me saying "Bring it on Baby!" Followed by doing my Victory Dance when the quake arrives.

Hang in there Amit, if there's a will, there's a way.

Love Ya.. Petra Challus

Roger Hunter said...

Amit;

Pay no attention to the crazy lady.

She's just mad because she failed the test and still can't pass it even though I accommodated all her objections.

Now she's talking about catalog errors as though they might be in her favor.

Anyway, for you to pass you would need 10 hits instead of 5 out of 21 predictions.

Roger

AMIT said...

Roger
hold it Roger.Thogh I am not aware of the said catlog error, We are here to discuss the matter in soft spoken words.Despite who is right and wrong. we need to be enough polite so that communication link is not boken alltogether
Regards
Amit

AMIT said...

Roger
By asking 10 hits out of 21 (7+ quakes with 3 day window period) you are asking
10/21 ie approx 1 in 2
whaere as average odds are 1 in 8 as mentioned above
Thus ,the requirement is 4 times better than average
Once you said 99% of average will be required where as now it is four times better than random selected dates
Amit

Roger Hunter said...

Amit;

Don't worry about it. Petra and I have a long history. She knows when I'm kidding.

You need to understand the statistics better. Especially as an engineer.

The 99% I use means that there is only 1 chance in 100 that the results are due to chance. It does not mean 99 out of 100 correct and with odds of 20.9%, 10 hits out of 21 tries would pass.

Roger

Anonymous said...

how could you do this?

Roger Hunter said...

Anonymous;

Do what?

Rogr

su said...

can u please explain ur thoery in simple words?

AMIT said...

hi
Please see my web page for theory in nutshell

http://earthquakeprediction.webs.com/

Amit