Weighting the Hot Spot

Using these three predictions:

Balloon Track for Windows                                Version 1.9.5
__________________________________________________________________________

                           Historical Predictions

Friday, November 03, 2006         EOSS-113          EOSS-113_pred_hist.csv
2:10:28 AM                         Windsor                Pred_history.txt
--------------------------------------------------------------------------

   Launch Site    |  Predict Date  |   Model   |  Flight Date   | Bear  | Range | Grid X | Grid Y |  Latitude |  Longitude 
------------------|----------------|-----------|----------------|-------|-------|--------|--------|-----------|------------
Windsor           | 11/01/06 01:47 |  GFS-Long | 11/11/06 12:00 |  67.2 | 141.0 |  107.8 |   78.5 |  41.23973 | -102.45873
Windsor           | 11/02/06 03:12 |  GFS-Long | 11/11/06 12:00 |  56.8 |  34.0 |    6.1 |   44.1 |  40.74196 | -104.41840
Windsor           | 11/03/06 02:07 |  GFS-Long | 11/11/06 12:00 | 104.3 |  71.4 |   46.7 |    7.4 |  40.21104 | -103.65031

For Hourly weighted averaging:

If the first prediction is 48 hours ago then:

The first prediction 48 hours ago carries a weight of one
A prediction 24 hours ago carries a weight of 24
And a prediction a few seconds ago carries a weight of 48.

The longer in the past the first prediction, the more pronounced the weighting will be. Well that's the case in all weighting models but the shorter the interval of weighting the more pronounced the focus shift is.

Weighted by hour