Monday, May 9, 2022

RISD May 2022 Elections Grade

So I made some predictions and the voters have decided. How did that go? Let's take a look at the RISD election outcomes. How did the voters do? How did I do?

In case you didn't read my predictions, you can find them here. As a reminder, my "predictions" were not scientific. I pretty much "sensed the wind" and applied some scenarios and took a shot.

District 5

Let's start off where I did the best: Richardson ISD Trustee in District 5.

That prediction was "correct" as the winning order was in my "Most Likely" column. I put "correct" in goofy quotes because I really was not predicting winners and losers as much as setting a level of expectation. To clarify, when I said "most likely" and I stated two possible outcomes, their order on the page is irrelevant. Because one was written before the other did not imply I thought one was more likely than then other.

The unofficial tally is:
Rachel McGowan - 2,962
Jan Stell - 2,221
Kile Brown - 624

I said, "the only difference is if the anti-CRT campaign has any traction or if there is a runoff at all." Indeed, the anti-CRT campaign (based on falsehoods in my opinion) did have traction. To quote the venerable H.L. Mencken,

The whole aim of practical politics is to keep the populace alarmed (and hence clamorous to be led to safety) by an endless series of hobgoblins, most of them imaginary.
This certainly describes the anti-CRT interjections into School Board politics

McGowan did have enough to avoid a run-off by having strong community support and by running a very strong campaign. This result is not surprising.

Now to where I didn't do as well.

District 2

In self-evaluating my own guesses I would call my District 2 prediction a "sort of" or "close but no cigar."

I stated, "The two wildcards which may determine the outcomes are the power of incumbency possibly benefiting Eron Linn, and the influence of the heavy PAC money spent in favor of Sherry Clemens.

The unofficial tally:
Sherry Clemens - 1,722
Vanessa Pacheco - 1,681
Eron Linn - 717

So the question of whether incumbency would boost Eron Linn was answered. The voters delivered a resounding "No" to that question.

With the second, the heavy outsider PAC money did effect the outcome but maybe not as much as it looks. Perhaps official results will yield a different take, but of the 14 precincts in District 2, Sherry Clemens won only 3. One of those was her home precinct. No candidate received a majority.

The eventual order (Clemens, Pacheco, Linn) was in my "Less Likely" column sandwiched between "Next Most Likely" and "Least Likely." The difference between that outcome and Pacheco,Clemens,Linn (which would have been in my "Most Likely" column) was only 41 votes. That is one vote less than 1% of the vote total. 

Did my predictions (i.e. guesses) take into account such a close race? No. I really wasn't putting that much effort into it. I said, "If Linn's incumbency falls flat then I see that it is possible that Clemens comes out on top but I think if that happens I think it is more likely that Pacheco does in that scenario."

Clearly only half of that occurred but it was less than a percent of total votes from being the case so I called that "close but no cigar." On the other hand, in both districts my "Least Likely" outcomes where nowhere close to happening so those were on the money.

I did say this was "...just as an experiment - partly amusing, partly serious - so I can be held accountable and hopefully gain bragging rights."

I can brag about District 5. Not so much in District 2, but I was less than a percent from being able to brag about that district too.

I will hopefully try to predict the run-off result. Watch here for that.


  1. Thanks for the effort. Your "system" needs work. You understood the factors at work here (power of incumbency, influence of PAC money, etc.). But when your District 2 "Most likely" predictions were for Pacheco or Linn to win, there's no way I would describe that as "close but no cigar." You need a mathematical scoring system, kind of like those scales at the fair have.

    1. Thanks for the comment. We will partly agree to disagree. Of course, one can always improve. however, less than one percent from "Most likely" (and is that not math and scoring), none of the "least likely" occurring, and another "Most likely" occurring is in the realm of "close but no cigar."

      If I was to get this result 20 times for 20 elections then the record would be at least 50% most likely occurring and 0% least likely occurring and the rest of differing probabilities. For a probabilistic prediction, that would not be so bad.

    2. 1% off would have been excellent. Your "Most likely" scenarios also contained Eron Linn coming out on top. You were 24% away from that. So do you average (1% + 24% / 2 = 12.5%) and say that's still "close but no cigar"? A system that allows you to predict two quite different outcomes as "most likely" and then claim success when one of the two comes close needs to be rethought.

    3. I look forward to your predictions with such a system.

    4. Someone would first have to define such a system. But it won't be me. I've never seen such a system for an election with two candidates (just ask Nate Silver), to say nothing of three candidates.