M5.1 SUCCESSFULLY PREDICTED by dsg on 4/8/2017

[Karma: 5 pts]

dsg

April 10, 2017 at 10:33 pm #1200

QW Required FORMAT for PROCLAIMING SUCCESS: In this order, post the following:
1) M5.1 SUCCESSFULLY PREDICTED by dsg on 4/8/2017
2) a link to all of your forecast postings that were active during that time.

New Caledonia:
Forecast: https://quakewatch.net/forums/topic/dsg-4817-318-pm-cst/
USGS Actual: https://earthquake.usgs.gov/earthquakes/eventpage/us10008ga9#map
Image: Google Drive Image – Print Screen of Distance

Score: 0

[Karma: 5 pts]

dsg

April 10, 2017 at 10:34 pm #1201

Please note that another member on this forum, CR was the first forum member to make an accurate forecast last week and I have sent him the link to this separate forum.

Score: 0

[Karma: 17 pts]

Management

April 14, 2017 at 9:06 pm #1244

Nice work. This type of prediction has some significance to it if it can be repeated. Have you considered focusing on this region? or are you open to looking elsewhere? Knowing when M5 events will happen is still relevant for sure. If you were focus there, and get 3 or 4 in a row with 4 or fewer days watch, it could be something. Upping the magnitude watches helps too.

Another way to pull it off would be to post a daily max magnitude for this region, which would allow us to analyze it that way as well.

Good job.

Score: 0

[Karma: 5 pts]

dsg

April 14, 2017 at 9:34 pm #1245

Thank you for the guidance. I have been searching for the best method to approach, and your insight helps to guide me.

I have done a great deal of research in the North and South Pacific for the last five years, as well as Africa. My ultimate goal was to refine the method I am using to get as narrow radius as possible, with the tools from the Disaster Prediction App and the QW and Space Weather News websites – and try to hit each of the major alert zones, and also chase down all tropical cyclones. My process is a slow and manual at the moment, using a variety of tools and maps/data sets across 7 platforms and some new technology. This one hit (and all the other misses) has helped me to reduce the actual time it takes for each one, and modify the formula with every error I have made.

I’ve been super close in geography on many, but have had excessive failure in magnitude estimation, so I am right in the heart of building a scale (from this New Caledonia model) to work with, that should ensure greater accuracy. One challenge – was the few I posted today – I lengthened the time-frame to 14 days, even though I interpreted the data to be just 48 hours or less away, and another, just 7 days away. Would you recommend that on those where the model suggests a certain timing, for me to call that out specifically – and still leave the main forecast/danger zone for the full week or two weeks? I like your idea of the daily max. Thanks to you guys for having patience with me while I stumble through trying to formulate a posting protocol that works well for communicating the variables.

Score: 0