Monday, July 4, 2016

The Upside of Down: Chapter 3

THE EXPERIMENTERS: Why There Are No Guarantees in Hollywood or Silicon Valley

A.K.A. Veruca Salt Discovers Confirmation Bias

I have a hypothesis that the real purpose of The Upside of Down is to retroactively justify McArdle's unearned professional success. "But you can't stand McArdle. Didn't you believe that anyway?" Yeah, but after three chapters I can start to prove it. The stated thesis of this particular codex is "How to fail well," but the chapters thus far have been only tangentially connected to that. On the other hand, every single one has served to address some blot on McArdle's own record, either by justifying it as something that happens or, as in this chapter, by disarming her critics.

We open with a little of that legerdemain that good motivational writers are known for. She describes a Hollywood disaster, a 90's epic with a ever-ballooning budget, numerous delays caused by the difficulty of shooting around water, and a big-name director who rapidly lost confidence in his own project. The outcome?

Titanic went on to become the top-grossing picture of all time


BOOM! You thought it was Waterworld, didn't you? Ha! Screw you, journalists who said the preview was tepid and that I don't know anything about economics!

...But yeah, Waterworld was a legendary flop, so what does this teach us? Simple: No one can know anything. Seriously, that's the thesis of this chapter, or at least part of it. Presumably, the lesson here is "...so don't be afraid to go out there and try new things," which is a little dubious. Given McArdle's track record, it seems more like an effort to dodge the consequences of all the false predictions she's made over the years, on Gulf War II and more economic crises than I care to recall. It's pretty convenient for an economist to decide that making mistakes is a learning experience, even a fake economist like McArdle. But as we'll be seeing shortly, some of her evidence actually blows back in her face.

McArdle starts off by citing the work of one Philip Tetlock, and this time I'm actually very familiar with him as Tetlock's research keeps turning up in books I've read. It was the heart of Future Babble, a book on why people make and believe in sweeping predictions, and also appeared in passing in The Invisible Gorilla in a section on confidence and the illusion of knowledge. But Tetlock has also written a few books on his own research, most recently Superforecasting in 2015. If you're interested, I'd suggest one of those books as McArdle's version is underwhelming and inaccurate.

Here's the overly truncated blogger version: Tetlock's experiment concerned expert predictions and was truly massive. It spanned nearly two decades (necessary to gather long-term accuracy), included thousands of experts making hundreds of predictions about major world events (starting in the 80's, when there was plenty of volatility) and tracking not only the individual predictions but the confidence behind them. In the end, the experts who made big, sweeping predictions and were sure of themselves ("hedgehogs") were the least accurate, while the ones who made moderate predictions and hedged their bets ("foxes") were the most accurate, though even the latter didn't do a great job and you could be nearly as accurate simply by assuming that tomorrow will look a lot like today.

McArdle...well, she adds to it a little:

Hedgehogs were more likely to be "very right" (think of Winston Churchill declaring Hitler's Germany an existential threat to Britain while Neville Chamberlain was declaring that he'd secured "Peace in Our Time"). But they were also more likely to be "very wrong" (think of Winston Churchill's disastrous decision as Chancellor of the Exchequer to return Britain to the gold standard after World War II...).
I predict that this post runs long.

Here McArdle has stumbled into the realm of confirmation bias, that nasty little cognitive blind spot that makes people more likely to remember successes than failures when it confirms something they wanted to believe anyway. Mix in a touch of the appeal to authority and you have a pretty solid explanation as to how even the worst of pundits can stick around year after year despite an impressive track record of faceplants. And McArdle freely admits that media and political figures benefit from making bold long-shot predictions - no one gets famous with a consistent history of accurate, middle-of-the-road forecasts, you get famous by making a flurry of huge, highly entertaining guesses and hoping that one or two of them are accurate enough that everyone remembers them.

Aside from making McArdle's entire profession look suspect, how does this advance her thesis? Well...it's not written anywhere in the text, but taking into account that this is a business book, one could argue that it's saying "you'll never know if your new product/service/strategy will succeed, no one can, so go for it anyway and learn how to lose with style." Without making a value judgment on the actual idea, I can say that it's internally consistent.

But I know McArdle, and I have an alternative explanation that ties this section in to the unstated thesis, but we'll get back to that. First, in the interest of wrapping this up in less time than the second chapter, I'll skip the in-depth and give a brief summary of the following seventeen pages. What McArdle does is detail a series of "experiments" intended to back up the above statement of uncertainty in much the same way as that opening laffer on Titanic and Waterworld. As follows:

  • A barely-relevant section on why people like certain cultural products, ultimately reaching the conclusion that no one knows but people like things that are already popular. That's based solely on personal choice, leaving out the structural reasons why popularity in the algorithm-driven internet tends to be a self-reinforcing loop (but that's a personal hobgoblin, even less relevant than the actual experiment).
  • A section on how welfare reform worked, despite what those experts said. Before you start sharpening your axes, McArdle - playing the role of the good Sensible Centrist - acknowledges that the improvement in the poverty rate during the Clinton years might - might - have had something to do with the generally strong economy. Maybe, but come on - we know what really got those lazy poors moving, right?
  • A naked bit of wingnut boob-bait on healthy food in schools, thank you Moochelle. I've been ignoring the endnotes until now, but this particular chunk rests entirely on two short articles, here and here. To paraphrase the person whose glowing praise adorns the back of the book, never has a case been made with such detail or such care.
  • A section on New Coke, because it is a business book. Can't talk about business without talking about New Coke, right?
Now, all of these examples, the warning to be cautious about big predictions, the criticism of excessive market testing - these are all consistent with a book advising board members and CEOs to acknowledge failure as an always present risk and to accept it rather than fearing it. However, it's not a really tight fit. It still meanders a lot, much like the first chapter did. You can see the partisan political nonsense creeping into it, and an overview of the chapter titles and endnotes suggest that there will be more. And above all, McArdle spends as much space cracking on people for being wrong (which obviously clashes with the stated thesis) as she does on giving business advice.

The unstated thesis is "Megan McArdle completely deserves her success despite what you dummies think." Does this chapter support that thesis?

Well, let's consider the ragging on experts with her half-understood version of Tetlock. Remember, this is a woman with a history of being really, really wrong. She began her glorious ascent into the media realm with a slate of predictions about Iraq that were as wrong as you get - she was one of those people who swore that the war would be nearly bloodless and would pay for itself. This is a woman who made up numbers because they felt right to her and then, when called out on it, claimed that the false number was a hypothetical. And while, owing to her shift into Brooksian Sensible Centrism, she's not making as many big impressive forecasts anymore (preferring to discuss safe topics like the culinary value of exotic salt), she's still dogged by critics who point out her many errors and question why she's received so many opportunities when people who were actually right about things continue to labor in obscurity.

Oh, and those people think they're so smart, don't they? Well, maybe I was wrong, but it's only because I was bold! And you experts were wrong, too - the research proves it! You don't know anything, you were just lucky! And success is systemic, so your theories were probably wrong anyway! You liberal types with your primitive hunter-gatherer fixed self-defeating mindsets, you'll never learn like I did! That's why I have a great job! Not because of my daddy's connections!

That's what I kept hearing as I read this chapter. Yes, you can absolutely make a case for everything in here being tied to the thesis as described in the jacket copy. But I can make a case of equal strength that this is our own grown-up Veruca Salt sticking her tongue out at those nasty bloggers and journalists and experts who won't let her sit at the big kid geopolitics table.

McArdle hasn't grown. I look at her, I see a woman who swears that she's learned from failure, even though she refuses to acknowledge even a fraction of her errors. Who attacks her critics by innuendo because she keeps failing on the facts. Who mistakes recklessness for boldness, and indecision for caution. For the love of God, this is a woman who argued that gun control was such a non-starter that we should instead stop mass shooters by teaching people - children, no less - to charge directly into gunfire and overwhelm the shooter with the weight of their bullet-riddled corpses. Check the dates - she wrote that while she was researching this book. Does that sound like someone who's learned from her mistakes?

Anyway, next chapter we (allegedly) get back to the thesis and to McArdle's advice. I'm thrilled.

No comments:

Post a Comment