| |
 |
|
|
Science Forum Index » Geology - Earthquakes Forum » Earthquake Dart Board final score
Page 1 of 1
|
| Author |
Message |
| Skywise |
 Posted: Tue Oct 16, 2007 12:38 am |
|
|
|
Guest
|
I quit updating my Earthquake Dart Board back in early March of this
year. The project lasted for just over 16 months with impressive
results.
Magnitude range: 4.00 - 10.00
Distance range: 0 - 1000 kilometers
Predictions: 5489 (total number of predictions made so far)
Prediction Hits: 3688 (number of predictions with at least one hit)
Total candidates: 18952 (total number of quakes within the magnitude
range)
Total hits: 17450 (number of candidate quakes within range of a
prediction)
Candidates missed: 1502 (number of candidate quakes missed)
Hit/Candidate Ratio: 92.075%
Hit/Prediction Ratio: 67.189% (predictions with hits vs total predictions)
To put it simply, 67% of my predictions successfully predicted
92% of the worlds earthquakes of magnitude 4.5 or above.
http://www.skywise711.com/quakes/EQDB/index.html
Brian
--
http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism
Seismic FAQ: http://www.skywise711.com/SeismicFAQ/SeismicFAQ.html
Quake "predictions": http://www.skywise711.com/quakes/EQDB/index.html
Sed quis custodiet ipsos Custodes? |
|
|
| Back to top |
|
| Thomas A. Russ |
| Posted: Tue Oct 16, 2007 3:51 pm |
|
|
|
Guest
|
Skywise <into@oblivion.nothing.com> writes:
Quote: I quit updating my Earthquake Dart Board back in early March of this
year. The project lasted for just over 16 months with impressive
results.
Magnitude range: 4.00 - 10.00
Distance range: 0 - 1000 kilometers
Predictions: 5489 (total number of predictions made so far)
Prediction Hits: 3688 (number of predictions with at least one hit)
Total candidates: 18952 (total number of quakes within the magnitude
range)
Total hits: 17450 (number of candidate quakes within range of a
prediction)
Candidates missed: 1502 (number of candidate quakes missed)
Hit/Candidate Ratio: 92.075%
Hit/Prediction Ratio: 67.189% (predictions with hits vs total predictions)
To put it simply, 67% of my predictions successfully predicted
92% of the worlds earthquakes of magnitude 4.5 or above.
But what's missing from this analysis is any comparison with the null
hypothesis. What we need to know is that if one were to randomly
generate time and location pairs covering the same time and area, what
sort of performance would that have?
Or is the name "dart board" supposed to mean that your experiment IS the
null hypothesis? Is this the result of a "pure chance" forecasting
method, or is there a claim that the forecasts are better than random
guessing?
--
Thomas A. Russ, USC/Information Sciences Institute |
|
|
| Back to top |
|
| Skywise |
| Posted: Tue Oct 16, 2007 11:11 pm |
|
|
|
Guest
|
tar@sevak.isi.edu (Thomas A. Russ) wrote in
news:ymizlyi3hr3.fsf@blackcat.isi.edu:
Quote: Or is the name "dart board" supposed to mean that your experiment IS the
null hypothesis? Is this the result of a "pure chance" forecasting
method, or is there a claim that the forecasts are better than random
guessing?
Basically....Yes.
So many predictors tout their enormous library of hits or their
percentages. My project shows quite plainly that those numbers
mean nothing without context.
Hmmmm....17,450 successfully predicted quakes in 16 months? How
come I'm not on C2C? I'll bet that's more than JOB and Deyo combined!
Brian
--
http://www.skywise711.com - Lasers, Seismology, Astronomy, Skepticism
Seismic FAQ: http://www.skywise711.com/SeismicFAQ/SeismicFAQ.html
Quake "predictions": http://www.skywise711.com/quakes/EQDB/index.html
Sed quis custodiet ipsos Custodes? |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Sun Nov 23, 2008 4:44 am
|
|