Can earthquakes be predicted?

Total Pageviews

Search This Blog

Powered By Blogger

Tuesday, October 08, 2013

Earthquake prediction evaluation and change in predictions

hi
  Time and again ,it is observed  that ,there are doubts about my predictions as to whether they are below averages or above. To facilitate ,the evaluation and to reduce the odds,hence forth I will be predicting quakes those are above 6.5
  Those who wish to check accuracy may please search this blog for " date hit prediction"
  Thus,there will be lesser  .dates of predictions.
 Amit

7 comments:

Roger Hunter said...

Amit;

How do you think that will help?

6.5+ has a probability too.

Roger

Roger Hunter said...

Amit;

First run on all your predictions (except most of 2013) taking mag and date into account show you are WAY worse than chance. I can't update my quake lists until NEIC gets back in business.

The problem is that you predicted too many small mags. Such quakes are very common so the probability is very high and missing one counts heavily against you whereas hitting one is worth very little.

Roger

AMIT said...

Roger
I have requested two things
1)If we run random dates ,one in how many trails we get the odds, as calculated by you
2)If we run random dates considareble time whether the results approch the theoritical odd
Thus we can test whether the odds (theoritical)are correct
Amit

Roger Hunter said...

Amit;

Life is not that simple. If you toss a coin 1000 times you will not always get 500 heads and 500 tails even though the odds are 0.500 for a fair coin.

The odds I use are not theoretical, they are observed and therefore accurate for the time period examined but that does not mean your results will be exactly as expected.

For that reason, I want results which far exceed the measured chance. And even if you get results significant at the 99% level, it just means that you're doing something that seems promising and deserves further study.

Roger

AMIT said...

Roger
I could not understand ,why the answers of my two points can not be given
I know only one thing..
If we run toss the coin considerable times ,the results should approach to the theoretical odds. If you toss coin more times ,the cumulative results should be going towards 0.5
If not, the theoretical odds are wrong
Similarly, if you run random dates on software and find out the quakes ,the results should also approach to your theoretical odds. If not there is something wrong with the theoretical odds.
Amit

Roger Hunter said...

Amit;

I just finished evaluating all your predictions from 2006 to Sep. 2013 using the odds on each prediction individually.

They were slightly below chance, with 63 hits out of over 200 predictions.

Roger

Roger Hunter said...

Amit;

I revised the prediction list for 2012 and 2013 as you suggested and reevaluated.

You're now exactly at chance level.

Not much help there.

Roger