hi
please see my prediction dated 25th Dcember 2010, where ,an earthquakes of 6.3 to 6.6 was predicted . Actual quake is 6.3 south of fiji Island ,to day, ie 28th december 2010( A date hit)
Also please see ,my web page where this was predicted (under -calender)
http://earthquakeprediction.webs.com/
regards
Amit
32 comments:
the probability of being correct in predicting earthquakes above 6.0 (inclussive of 6.0R) is 50% approx. when one uses 2 day window. When one uses 3 day window it is a certainty.
So predicting 6R earthquakes is not of added value.
Remi
PS Serious stats are needed and I see no thought being give on them in here AT ALL. Sad really.
Annymous
You have not given some random dates,to prove what you say ,is random ,and any body can do it.
I have, one more point to say
If both of us give probable dates ,say 20 odd ,in the year 2011, for earthquakes ,greater than 7,what are the odds?
Are you ready to take up this task, and give 20 odd dates ,for quakes 7+( for 2011),for comparison
I think , the issue of certainty ,will not arise then.
regards
Amit
Yes I will take up your challenge. By the way, I am interested in your method but I expected from you more serious work on it.
I will post 20 days for 2011 for >7R
I expect that you will also post your 20 days simultaneously with me. How shall we do this?
Maybe I will post mine first and then you post yours afterwards I am happy with that. Give me a day to sort it.
Remi
Amit;
Based on the 2010 performance the odds on getting a hit with a random 3 day window are 0.686, not 1.000.
Roger
Anonymous
Thank you very much ,for your response.
I will post my dtaes(20 odd) for 7+ quakes(with 3 day window),for 2011 on 9th January 2011,sunday, between 0600 Utc to 1800 UTC
Thanks again
Amit
Roger
Thanks roger
Amit
Looking in USGS Stats there are 150 on average >6R quakes per annum.
122 3day windows per annum. 150/122 quakes per year. >1. That maybe simple but good enuf.
The average number of 7+ per annum is 15. That is certainly more valuable as we have to pick 15 in 122 windows 15/122 = 12.3% random excluding clustering which we will ignore as this is frills for now.
Hi Amit,
Here we have to clarify. I said that the chances of >6R quake is too good globally to require a method. You challenged and I accept this.
IF you say now 20 days for 7+, the chances randomly to get it right are less as I and Roget pointed out. I will take the 7+ challenge as you suggested and I will soon give 20 days for 2011. But what are we going to compare? Random versus your method?? We will compare 2 sets of data as to which one is best by comparison to the dates of the exact 7+ quakes. OK? My method versus your method. I says this to clarify that we agree that predicting 6R quakes is not a challenge and 7R is a better challenge. Well, having agreed that 7R is the quakes to look for I hope your dates win compared to mine thats the idea. However this is something you could have done using past dates! That is why I am amazed you did not do any stats.
I have a request.
Does anybody know where to find a listing on the web the daily speed of planets. We know that perihelion is fastest but what is the speed of a planet daily is what I want. Is there a URL for it?
Thanks
REMI
Anonymous
Please visit my web page
http://earthquakeprediction.webs.com/ and go to links. There is a link for planet epemeris-daily positions. yu can get normal speed rom there . Fpr ready reference nrmal speed of planets are
1)Mars---31 Minutes
2)Mrecurry----59 minuts
3)Jupiter-----5 minutes
4) venus--------59 minutes
Saturn------- 2 minutes
regards
Amit
Anonymous
results of 7+ quake prediction for the year is already posted on my web page
http://earthquakeprediction.webs.com/ under News category
If we consider major earthquakes,7+,for prediction accuracy for the year 2010,it would be like this
1) There are 45 dates predicted for 6+ quakes.(some of them-4- stated as 7+ also)
2)There were 22 quakes of 7+ in year 2010 (till 26th December 2010)
3)7+ quakes occured in window period of prediction are 12.( out of which 4 were stated as 7+)if we do not consider clusters of 7+ quakes ,it is 10 hits in one year (2010)
what are the odds of getting ten 7+ quakes hit in window period of 3 day,in a year?
Now ,if we consider 8+ quakes,there is only one 8+ quakes on 27th February (chile).The date predicted was 28th february 2010. so ,it is in window period.
what is odd of getting 8+ quake in 3 day window?
regards
Amit
To help you out on the numbers. Assuming binomial distribution, or Poisson but the difference is small between them, (and a problem here is that the events are sequential and not simultaneously occurring, so much better to use probabilistic approach, but that is another discussion) then for the 7+ case, where p=0.123 per year, 3 day slot,
then
P(0 success)= 7.24%
P(1 success) = 20.3%
P(2 successes) = 27%
P(3 successes) = 22.7%
P(4 successes) = 13.58%
etc etc
So if we get 4 correct 7+ in a year this has prob of 13.58% according to a Binomial Distribution...Hope this helps.
Remi
For the 8+ Quakes where we get 1 a year,
p=1/122= 0.8%
In 43 trials P(1 success) = 35%
[P(success) = C(42,1)*p*(1-p)]
C()=Combinations
OK?
AMIT,
I visited your links but the Ephymeris link does not have the planet speed. Can you post the exact link for me please...How do you find the planet speed?
Reg.
REMI
Stats:
You got 12 hits out of 22 possible.
P(12 hits)=22C12*p^12*(1-p)^10
p=22/122=0.1803
P(12 hits) = 0.01%
;-)
Hi Amit,
In 2010 we have had 24 7+ quakes.
You have predicted 9 from a sample of 43 published dates in January 2010. (i.e. 37.5%). .... The window was 3 days wide.
Remi.
sorry 21 7+
correction...
Anonymous Anonymous said...
For the 8+ Quakes where we get 1 a year,
p=1/122= 0.8%
In 43 trials P(1 success) = 24.5%
[P(success) = C(42,1)*p*(1-p)^42]
for p=0.8% we have P(1)=24.5%....
Anonymous
We can not say it as 43 trials.This is because my magnitude predicted was for +or- 0.3. Hence only those which I have predicted 6.7 and above
pl see my prediction at 10th January 2010 blog
Amit
Amit,
It is about time you do some work and stats instead of me.
;-)
Happy New Year.
Anonymous
Yes sir. Happy New year
Amit Dave
Remi
The ephemeris link stated will give daily position of the planets. If we deduct the degree /minutes of the previous date from todays position ,we will get the daily speed of that planet
Amit
hi
1)For the comming year 2011, I will only predict the earthquakes those are 7+.
2)Few dates ,I will mention as 7.5+
3) The window period will be + or - 1 day,ie total 3 day window
4)The time given will be in UTC. and window period will be + or - 36 Hrs (ie 3 day window)
5) Te magnitude window will be + or -.03
6) Place will not be predicted
7) Above all , The stats and analysis will be done by all experts in the field. Let the dates speak
regards
Amit
hi
correction for sr number (5)
The magntude window will be + or - 0.3
Amit
Amit;
For 2010, if I check mags using your +0.3,-0.3 range, you only get 12 hits out of 43 predictions
Roger
Amit;
I tried the 2011 rules on your 2010 predictions and out of 16 mag 7 predictions you got only 3 hits.
That 36 hour idea will cause me problems as it will lap over into the second day before or after the predicted date depending on whether the time is before or after noon.
Not insurmountable, just a nuisance.
Roger
Roger,
if you got 3 hits out of 16 that is about 19% less than random 24.5% is this right?
Anonymous;
I get 0.230 for the odds on getting a 7.0 +-.3 quake in a 3 day window.
If that's correct, getting 3 hits with 16 tries is just about right.
Roger
Thanks Roger,
would be interested if it is better to use information theory in evaluating the 'added value' of a prediction. You have some paper(s) I am sure. let me know if you are willing to share them.
Regards,
Remi
Remi;
No, actually I don't have any papers on it.
Roger
Post a Comment