by Mish |
That blatant bias shows in projections.
Nate Silver and his “expert panel” has continually understated Trump’s odds of winning.
In January, Silver assessed Trump’s odds of winning in the high “single digits”. Last August it was something like 3%.
Silver promotes either biased BS or poor analysis (take your pick), but at least his “expert panel” posts numbers for everyone to mock (and I have done so correctly for months on end).
In contrast, mainstream media promotes the idea that Trump is an underdog to win the Republican nomination without posting any analysis or numbers at all.
What brought about the above rant was the following snip from the Financial Times article Donald Trump Sweeps New York Republican Primary.
Mr Trump effectively closed off the path for Mr Cruz to win the 1,237 delegates needed for the nomination on the first ballot in Cleveland. But the better than expected result does not significantly alter the trajectory of the Republican race, which is still more likely than not to end with a contested convention.
Say What?
Financial Times, excuse me for asking, but can I please see your numbers?
Here are mine, updated as of tonight, showing my predictions vs.s Silver’s predictions (his much revised, mine barely).
Starting tonight, let’s track Silver’s projections at the time he made them vs. my projections made at the same time.
Mish vs. Silver
Silver initially projected 71 New York delegates to Trump. When he predicted 71, I predicted 86.
I was too pessimistic. Trump is going to win 90 or 91.
Running Score Wyoming
Silver projected Trump would win 1 Wyoming delegate
I predicted 0.
Trump won 0.
Running Score New York
Silver originally projected 71.
I originally projected 86.
Trump will win 90 or 91.
Indiana
Silver originally projected 37
I originally projected 45
Silver lowered his projection to 9
I kept my projection at 45
Indiana may very well determine the outcome of this race. If Trump wins Indiana, it’s likely all over.
Trump will win the nomination if he wins Indiana unless there is a last minute rule change.
California
As recently as April 15, Nate Sllver’s personal projection was that Cruz had a 60% chance of winning California. One day later Silver reversed.
How pathetic is that?
Meanwhile his “Polls Only” forecast looks like this.
The Nomination
My projection of 1224 is 13 short. But there are 125 uncommitted delegates. 13 is not a lot to ask, and Trump is likely to get that number from Pennsylvania alone.
If the Financial Times wishes to state it is “still more likely than not to end with a contested convention” then I demand to see the Financial Time’s math.
Put up or shut up. At least Silver had the decency to make a fool of himself with his continuous under-assessment of Trump’s odds.
I am arguably in the same boat with my statement “Trump Will Win Indiana”, but I was never so silly or so biased as to believe Trump had 3% chances in August or single digit chances in January.
Silver is totally clueless on judging human sentiment should it veer outside historical norms.
That said, his final numbers, continually revised, will most assuredly beat my projections made months ago. For that he will be labeled a genius.
As for me, heaven help me if I am wrong about Indiana.
Mike “Mish” Shedlock
