Prediction ( J42 ) 5 / 11 / 2016
#11
Hi Roger

This post, like the event is old news.  I was just following up on what I had posted, I didn't have the slightest notion it would be accepted, so don't worry about it.  Technically speaking, there has not been a mag 5 quake in that region of Samar for 11 months.  So if you saw this from my point of view, an event occurs 19 days after I mentioned where there would be one, and again it is swept away in statistics.   Consequently, you end up with the wrong end of my frustration.

I'm not concerned about a comprehensive list, statistical analysis " will " show I am wrong, which is OK.   I have come to terms with the fact that testing physics with geological science will never work.  Movements of the plates will always produce earthquakes, that's guaranteed.  So called signals from anomalies are only as reliable as the energy detected from pre-seismic events. You may have multiple detection's, but it only needs one big event to suppress the energy in the remaining anomalies, thus consistency is very difficult to achieve ... something that will have little understanding here.   I recently posted a block of 25 predictions, I got the first eight in a row correct, the last being New Zealand.  The remaining produced no results, except for the 7.9 Solomon event, which according to Chris is a little out of the ordinary with regards the time period, of two other similar mag quakes.  Statistically, the numbers fail me, so it is akin to pulling numbers out of a hat. Scientifically, if I used the same method of detection for all predictions, why did the rest fail after the NZ event ?  Chance or no ... recent results in however light you wish to view them, show that I am beyond chasing transmitter towers or electrical sub stations.  As you said yourself, the only way I can accomplish anything here is to get most or all ... hope you mean that !

All the best

Duffy




Reply
#12
(12-22-2016, 03:53 PM)Duffy Wrote: Hi Duffy

Quote:This post, like the event is old news.  I was just following up on what I had posted, I didn't have the slightest notion it would be accepted, so don't worry about it.  Technically speaking, there has not been a mag 5 quake in that region of Samar for 11 months.  So if you saw this from my point of view, an event occurs 19 days after I mentioned where there would be one, and again it is swept away in statistics.   Consequently, you end up with the wrong end of my frustration.

You are mistaken. The methods I use check the seismicity of your area as part of the evaluation process so your rare event would have received a high score.

Quote:I'm not concerned about a comprehensive list, statistical analysis " will " show I am wrong, which is OK.   I have come to terms with the fact that testing physics with geological science will never work.

Wrong again. All that's required here is that you are correct more often than chance would allow. Much more often; it's a tough test.

Quote: Movements of the plates will always produce earthquakes, that's guaranteed.  So called signals from anomalies are only as reliable as the energy detected from pre-seismic events. You may have multiple detection's, but it only needs one big event to suppress the energy in the remaining anomalies, thus consistency is very difficult to achieve ... something that will have little understanding here.   I recently posted a block of 25 predictions, I got the first eight in a row correct, the last being New Zealand.  The remaining produced no results, except for the 7.9 Solomon event, which according to Chris is a little out of the ordinary with regards the time period, of two other similar mag quakes.  Statistically, the numbers fail me, so it is akin to pulling numbers out of a hat. Scientifically, if I used the same method of detection for all predictions, why did the rest fail after the NZ event ?  Chance or no ... recent results in however light you wish to view them, show that I am beyond chasing transmitter towers or electrical sub stations.  As you said yourself, the only way I can accomplish anything here is to get most or all ... hope you mean that !

Oh I do mean that. I've been hoping to find someone who could do it, no matter what the method might be.

Quote:All the best

And the same to you. Remember, I'm not the enemy.

Roger




Reply
#13
Roger

I thought I would give you something for Christmas, in thanks for your encouraging words shall we say !  

Azores Cape St. Vincent Ridge ... 15' 54' W by 35' 56' N +/- 1' ... 4.5 - 5.5 ... 30 days from time of this post

This is basically a small test to determine the value of individual history of a location, against statistical analysis of multitudes.  According to the records I have sourced, the last two events above mag 4 in this region happened in August this year. They occurred at 11 and 12 degrees longitude respectively.  Last event before this was a 4.4 in March 2015, it is also the last event on 15 degrees west, related to said ridge.  So from my perspective, using a 30 day window at this location, would not favour certainty of success. History here speaks for itself, period of event seems to justify period of window.   Using the same window in the Kermadec's works against me, only because it is a very active region ... individual location doesn't count here.  

Because I have addressed this location, and should the event occur ... it should imply a certain skill factor was involved in it's determination, weather the data is understood or not, specifically because of it's history.  Circumstance cheats me any accreditation  because I have to track the multitudes to find true events, and in doing so, it lowers my percentage of success. I am currently working on a new system which I hope will tip the balance in my favour, should be online in spring.

Signals from this location are pretty strong and constant, so I figure 90% on this one, the down side is, this could be an event in the Atlantic or an " actual " event in Iceland !!

I know your not my enemy, I wouldn't be giving a gift if you were ... hope you don't have to wait 30 days to unwrap it !



Duffy  (22:38 ut 22nd Dec)




Reply
#14
(12-22-2016, 10:38 PM)Duffy Wrote: Roger

I thought I would give you something for Christmas, in thanks for your encouraging words shall we say !  

Azores Cape St. Vincent Ridge ... 15' 54' W by 35' 56' N +/- 1' ... 4.5 - 5.5 ... 30 days from time of this post

This is basically a small test to determine the value of individual history of a location, against statistical analysis of multitudes.  According to the records I have sourced, the last two events above mag 4 in this region happened in August this year. They occurred at 11 and 12 degrees longitude respectively.  Last event before this was a 4.4 in March 2015, it is also the last event on 15 degrees west, related to said ridge.  So from my perspective, using a 30 day window at this location, would not favour certainty of success. History here speaks for itself, period of event seems to justify period of window.   Using the same window in the Kermadec's works against me, only because it is a very active region ... individual location doesn't count here.  

Because I have addressed this location, and should the event occur ... it should imply a certain skill factor was involved in it's determination, weather the data is understood or not, specifically because of it's history.  Circumstance cheats me any accreditation  because I have to track the multitudes to find true events, and in doing so, it lowers my percentage of success. I am currently working on a new system which I hope will tip the balance in my favour, should be online in spring.

Signals from this location are pretty strong and constant, so I figure 90% on this one, the down side is, this could be an event in the Atlantic or an " actual " event in Iceland !!

I know your not my enemy, I wouldn't be giving a gift if you were ... hope you don't have to wait 30 days to unwrap it !
--------------------------------------------------------------------------------------------------------------------------------------------------
Thanks for the kind words Duffy, I appreciate them. Most predictors hate to be evaluated and it affects the evaluator too.

a quick search of that tiny area for quakes in the 3.0 to 6.5 range turned up 18 cases from 1988 to 2008 so if you hit this one I'll definitely be interested but one is just a fluke - maybe. You'll need many more that good to establish credibility.

Roger




Reply
#15
(12-22-2016, 11:08 PM)Roger Hunter Wrote:
(12-22-2016, 10:38 PM)Duffy Wrote: Roger

I thought I would give you something for Christmas, in thanks for your encouraging words shall we say !  

Azores Cape St. Vincent Ridge ... 15' 54' W by 35' 56' N +/- 1' ... 4.5 - 5.5 ... 30 days from time of this post

This is basically a small test to determine the value of individual history of a location, against statistical analysis of multitudes.  According to the records I have sourced, the last two events above mag 4 in this region happened in August this year. They occurred at 11 and 12 degrees longitude respectively.  Last event before this was a 4.4 in March 2015, it is also the last event on 15 degrees west, related to said ridge.  So from my perspective, using a 30 day window at this location, would not favour certainty of success. History here speaks for itself, period of event seems to justify period of window.   Using the same window in the Kermadec's works against me, only because it is a very active region ... individual location doesn't count here.  

Because I have addressed this location, and should the event occur ... it should imply a certain skill factor was involved in it's determination, weather the data is understood or not, specifically because of it's history.  Circumstance cheats me any accreditation  because I have to track the multitudes to find true events, and in doing so, it lowers my percentage of success. I am currently working on a new system which I hope will tip the balance in my favour, should be online in spring.

Signals from this location are pretty strong and constant, so I figure 90% on this one, the down side is, this could be an event in the Atlantic or an " actual " event in Iceland !!

I know your not my enemy, I wouldn't be giving a gift if you were ... hope you don't have to wait 30 days to unwrap it !
--------------------------------------------------------------------------------------------------------------------------------------------------
Thanks for the kind words Duffy, I appreciate them. Most predictors hate to be evaluated and it affects the evaluator too.

a quick search of that tiny area for quakes in the 3.0 to 6.5 range turned up 18 cases from 1988 to 2008 so if you hit this one I'll definitely be interested but one is just a fluke - maybe. You'll need many more that good to establish credibility.

Roger
M 4.3 Azores-Cape St. Vincent Ridge 07:01:00 ut 6th March ... 16' 06' W - 37' 18' N ... 43 days after expiry

Roger;

Happy belated Christmas ... for what it's worth!

Duffy




Reply
#16
(03-06-2017, 10:23 AM)Duffy Wrote:
(12-22-2016, 11:08 PM)Roger Hunter Wrote:
(12-22-2016, 10:38 PM)Duffy Wrote: Roger

I thought I would give you something for Christmas, in thanks for your encouraging words shall we say !  

Azores Cape St. Vincent Ridge ... 15' 54' W by 35' 56' N +/- 1' ... 4.5 - 5.5 ... 30 days from time of this post

This is basically a small test to determine the value of individual history of a location, against statistical analysis of multitudes.  According to the records I have sourced, the last two events above mag 4 in this region happened in August this year. They occurred at 11 and 12 degrees longitude respectively.  Last event before this was a 4.4 in March 2015, it is also the last event on 15 degrees west, related to said ridge.  So from my perspective, using a 30 day window at this location, would not favour certainty of success. History here speaks for itself, period of event seems to justify period of window.   Using the same window in the Kermadec's works against me, only because it is a very active region ... individual location doesn't count here.  

Because I have addressed this location, and should the event occur ... it should imply a certain skill factor was involved in it's determination, weather the data is understood or not, specifically because of it's history.  Circumstance cheats me any accreditation  because I have to track the multitudes to find true events, and in doing so, it lowers my percentage of success. I am currently working on a new system which I hope will tip the balance in my favour, should be online in spring.

Signals from this location are pretty strong and constant, so I figure 90% on this one, the down side is, this could be an event in the Atlantic or an " actual " event in Iceland !!

I know your not my enemy, I wouldn't be giving a gift if you were ... hope you don't have to wait 30 days to unwrap it !
--------------------------------------------------------------------------------------------------------------------------------------------------
Thanks for the kind words Duffy, I appreciate them. Most predictors hate to be evaluated and it affects the evaluator too.

a quick search of that tiny area for quakes in the 3.0 to 6.5 range turned up 18 cases from 1988 to 2008 so if you hit this one I'll definitely be interested but one is just a fluke - maybe. You'll need many more that good to establish credibility.

Roger
M 4.3 Azores-Cape St. Vincent Ridge 07:01:00 ut 6th March ... 16' 06' W - 37' 18' N ... 43 days after expiry

Roger;

Happy belated Christmas ... for what it's worth!

Duffy

Duffy;

And the same to you.

But one hit (or miss) proves nothing. Any odds can be beaten - by chance - but not repeatedly.

Roger




Reply
#17
Roger;

According to what you posted, nothing has occurred here since 2008, and I am 43 days late.  

If I predict Sumatra, I gain no credit because there is one a month in Sumatra.

If I predict the Kermadec's, I gain no credit because there is one a month in the Kermadec's.

If I predict a location void of activity for nine years, I gain no credit because it is a one hit wonder.

I don't predict anymore because there is "nothing to gain" from it ... I just experiment!


Duffy




Reply
#18
(03-06-2017, 03:52 PM)Duffy Wrote: Roger;

According to what you posted, nothing has occurred here since 2008, and I am 43 days late.  

If I predict Sumatra, I gain no credit because there is one a month in Sumatra.

If I predict the Kermadec's, I gain no credit because there is one a month in the Kermadec's.

If I predict a location void of activity for nine years, I gain no credit because it is a one hit wonder.

I don't predict anymore because there is "nothing to gain" from it ... I just experiment!


Duffy

Duffy;

I'm sorry you don't like the rules but I didn't invent them.

Part of the problem is your 30 day window. It slants the odds against you.

One hit wonder, yes but if you do it repeatedly it's worth a lot.

Roger




Reply
#19
(03-06-2017, 07:08 PM)Roger Hunter Wrote:
(03-06-2017, 03:52 PM)Duffy Wrote: Roger;

According to what you posted, nothing has occurred here since 2008, and I am 43 days late.  

If I predict Sumatra, I gain no credit because there is one a month in Sumatra.

If I predict the Kermadec's, I gain no credit because there is one a month in the Kermadec's.

If I predict a location void of activity for nine years, I gain no credit because it is a one hit wonder.

I don't predict anymore because there is "nothing to gain" from it ... I just experiment!


Duffy

Duffy;

I'm sorry you don't like the rules but I didn't invent them.

Part of the problem is your 30 day window. It slants the odds against you.

One hit wonder, yes but if you do it repeatedly it's worth a lot.

Roger

Roger;

I'm not blaming you, and no I don't like the rules because they don't leave any room (or time) to develop new techniques.  I gave my opinion on why individual location history should be taken into account and that's all I can do.  If you want to brush this one under the rug like it was a 2.5 in Turkey, then that's your prerogative to do so ... your just following the rules right! 

If you want me to do this repeatedly ( which may not be as crazy as it sounds)  I need a 30 day window, and a 10 degree corridor to monitor ... bend the rules a bit, and put some incentive back into prediction !

Duffy




Reply
#20
(03-06-2017, 09:51 PM)Duffy Wrote:
(03-06-2017, 07:08 PM)Roger Hunter Wrote:
(03-06-2017, 03:52 PM)Duffy Wrote: Roger;

According to what you posted, nothing has occurred here since 2008, and I am 43 days late.  

If I predict Sumatra, I gain no credit because there is one a month in Sumatra.

If I predict the Kermadec's, I gain no credit because there is one a month in the Kermadec's.

If I predict a location void of activity for nine years, I gain no credit because it is a one hit wonder.

I don't predict anymore because there is "nothing to gain" from it ... I just experiment!


Duffy

Duffy;

I'm sorry you don't like the rules but I didn't invent them.

Part of the problem is your 30 day window. It slants the odds against you.

One hit wonder, yes but if you do it repeatedly it's worth a lot.

Roger

Roger;

I'm not blaming you, and no I don't like the rules because they don't leave any room (or time) to develop new techniques.  I gave my opinion on why individual location history should be taken into account and that's all I can do.  If you want to brush this one under the rug like it was a 2.5 in Turkey, then that's your prerogative to do so ... your just following the rules right! 

If you want me to do this repeatedly ( which may not be as crazy as it sounds)  I need a 30 day window, and a 10 degree corridor to monitor ... bend the rules a bit, and put some incentive back into prediction !

Duffy
Duffy;

Good news! I finally found the error in my program, a trivial subroutine to convert decimal degrees to hours and minutes wasn't doing what it should.

Now I can compute RA and dec for sun and moon for any date and time between 1900 and 2100.

Not that significant since the NEIC quake catalog starts in 1973 and isn't all that complete for smaller quakes.

Still, we do what we can with what we have. I propose to take one day per week, from 1950 to 2010, get sun and moon locations, then look for quakes in a 30 day window within a 90 degree distance in longitude +/- 1 degree.

Do you agree with these parameters? Dec is not being used here, just RA as longitude.

Roger




Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)