On wrong turns

For the past couple of months, my discussion of the Saskatchewan NDP leadership race has featured charts with my guesses as to the outcome - the last of which can be found here. In the absence of detailed polling data or other information providing a direct indication of voter intention, I've tried my best to make predictions based on the factors that seemed to me to be at play. But in retrospect, there are a couple of areas where my assumptions were apparently off the mark - and I'll take a few minutes to point those out.

The first and most obvious is my relative treatment of the candidacies of Yens Pedersen and Deb Higgins. In Pedersen's case, I underestimated the strength of his campaign, focusing too much his comparative lack of public endorsements when there were probably signs of solid organization that could have been given more weight, particularly the logic underlying his phone poll process.

In Higgins' case, I likely did just the opposite, overemphasizing her positive moments due to her endorsement base and giving plenty of weight to media narratives which in retrospect were likely based more on her name recognition than on anything happening during the course of the campaign. And the result was my regular ranking of Pedersen as far more likely to finish last than Higgins (contrary to the final results this weekend). So my apologies to Yens for missing the relative strength of his campaign - and hopefully he can take some small moral victory in having proven one self-appointed pundit wrong.

The second broad area has to do with the relative likelihood of a first-ballot win for Dwain Lingenfelter and his prospects for victory later on. In retrospect my analysis wasn't far from the mark in placing the two most likely results as a Lingenfelter first-ballot victory or a final-ballot showdown again Meili, with little to choose between the relative likelihood of the two. But I likely didn't account enough for the chances that a Lingenfelter first-ballot lead would get him within a small number of votes of the 50% mark such as to make a later-ballot win difficult for his competitors as proved to be the case.

I'll leave the personal navel-gazing at that for now, and hopefully learn a few lessons to be applied to future analysis. But thanks to all who have taken the time to read my off-the-cuff predictions - and I'll try to be closer to the mark next time.