J.R. (who might or might not be from the South, according to his sig), with some other Kossacks agreeing with him, suggested an alternate way of calculating the delegate counts based on poll results.
Their thesis is that very few people go and vote "undecided/don't know" in the actual primaries. Therefore it would be fair to look at the percentage of the poll responses that were not undecided.
Therefore if Clark has 13% in Alabama, with 38% undecided, then his percentage w/ the undecideds removed is 13/(100-38) * 100, or 21.
This new set of numbers is shown in the orange bar, underneath the delegate totals computed the old way. I'll leave these as an alternate stat for now, until I get some more feedback.
Using the new stat, Clark drops back to 2nd place, w/ 1491 (from 1774) to Dean's 1668 (from 1558). Those two remain far ahead of 3rd place Lieberman, who has 661 (a gain of 50 delegates from the count that includes undecideds.) Kerry and Gephardt pick up 40 and 89 respectively using the new stat.
Click here to view the latest numbers (225 KBytes)
Now, in case anyone else wants to play w/ these numbers, you can right click on the link to the latest numbers and choose SaveTarget as. Save to your desktop, then open with Microsoft Excel 2000. Select all, right click, and choose Unhide. Feel free to post a comment or E-mail me if you want any explanation of what the different columns and formulae mean.
Click here for previous diary entries.
And thanks for the brilliant feedback! If anyone can come up w/ a logical way to use neighboring state numbers for states w/o polls instead of national #s, that would be very cool... please let me know.