Supported by Greenhaven Road Capital, finding value off the beaten path.
Update: This is a two-parter. First, the update from Falk on Wharton Moneyball and below from his podcast with Zach Lowe.
Getting lucky. When asked how he got started in analytics Falk said:
“The short answer is right place, right time.”
“But also the right time of my life. At the time I was in high school and college and could explore these things. I could volunteer, I didn’t have to take a paying job. It kind of all just came together.”
Like Ezra Klein, Yvon Chouinard, or the Wright brothers, Falk didn’t have demands on his time. This was Malcolm Gladwell’s point when he wrote about ten-thousand hours. If you spend the time becoming great at something that’s time you don’t spend doing something else.
About working for Sam Hinkie.
“Sam very much valued a different kind of thinking and having different backgrounds and we had a diversity of backgrounds.”
“(The range of problems we thought about) was as wide as you could imagine. That was one of the beauties of working for Sam. It was about a method of thought more than anything and that method of thought could be applied to any problem you faced in basketball operations.”
The meta game. In my post about Sam Hinkie I noted that poor communication was Hinkie’s Achilles Heel. Josh Brown advised financial advisors, “If you’re just telling a client, ‘Shut up I got this,’ you’re not going to be the client’s advisor for a long period of time.” Falk said:
“I think there are things that could have been done from a PR or media standpoint to make it slightly more palatable.”
“Sam and I would often talk about the metagame. There’s the game itself which you might be able to optimize but there’s this game outside the game which is important to optimize as well. (Selling an unconventional strategy to others.) Yep, and that impacts whether your strategy works.”
Falk points out that if it were a video game you might do one thing, but in reality, you’d do another. It’s why we pointed out in Rory Week that spreadsheets don’t have elbow room.
… …. …
Ben Falk joined Zach Lowe to talk about constructing NBA teams and it was excellent. I should have expected that. Falk worked with Sam Hinkie, who I wrote a post/love letter about. Lowe and Falk talk about a lot of ideas we’ve covered here, so let’s rapid-fire this.
Stakeholders matter. If your boss says ‘do X,’ it precludes doing Y and Z. If your boss says, ‘do what’s right,’ you can do X, Y, or Z. Team owners that say to ‘win now’ remove Y and Z.
“(This can be) a mandate that comes down from ownership, ‘I want you to do both. To win now and win in the future.’ It’s an incredibly competitive league, that’s very difficult to do. You need to get incredibly lucky to pull that off so you have to be willing to strategize in such a way as ask, ‘What are we willing to give up?'”
Investors see this with capital withdrawals. Businesses see it with customer departures. The best teams, investors, and businesses filter in the right stakeholders from the beginning. General Managers that want to rebuild a certain way will try to find opportunities where that ‘way’ aligns with the ownership. If you can, as Dan Egan says, attract wonderfully boring customers to your wonderfully boring company, you’ve got a good match.
Conditions matter. As Warren Buffett wrote, “the rise and fall of the tide is hardly something for the duck to quack about.” Investors admit that if you did well in a bull market all you really know is that you did well in a bull market. The same holds for sports.
“Early in the season, (during Falk’s stint in Portland) we had gotten blown out and everyone asked, ‘Is this a bad sign for the Blazers?’ It turned out, it was a good sign for the Suns.”
Mike Lombardi says this riddle exists in the NFL too. Did your team play well or did the other team play poorly?
Non-linearity. Lowe and Falk talk about many individual players and Zach said, “I think on a better team where he had to do less heavy lifting, Devin Booker’s shooting would really sing.” Lombardi recalled, “Scotty Williams looks great when he plays twenty-eight minutes for the Bulls but when he plays thirty-eight minutes he sucks.”
Lowe admits that this might happen to Kristaps Porzingis, “I don’t know that any projection systems factored in the enormous damage of having an atrocious point guard rotation. Getting into the offense is going to be a struggle. God bless Porzingas…but I worry if he can maintain this level of efficiency carrying that kind of load.”
Nassim Taleb likes to point out that snow resorts operate similarly. No snow is bad and so more is good – up to a point. Too much snow falls and no one can get to the resort.
You know I’m all about that base, bout that base, bout that base rate. “If you do a search of all rookies who are 18, 19, or 20,” said Lowe, “and played X amount of minutes and had a PER of less than 7.5 the results are not encouraging.” True, adds Falk, who said that you had to think probabilistically about outcomes. And “as soon as you get more numbers you should be updating those beliefs.”
Base rates are incredibly helpful to figure out what normally happens here. Hinkie said this in his podcast with Lowe:
“One of my big things is to survey base rates any time you can. You go back to facts whenever you can. You go back to finding what has worked. Often things have changed …but at the same time you should be staring with some base rate. What does this look like?”
Better predictions. In projecting a player, Falk said to think about things that can change and things that can’t.
“You try to think about things that are predictive of future ability and what are the things that can change.”
“Basketball IQ and understanding of the game doesn’t change as easily as people think. You can learn a little bit but if you come in not being able to pick up the nuances of the game it’s hard to change that.”
“This is one of the reasons shooting may be overrated going into the draft…but it’s easier to change than height or basketball IQ. It’s harder to find a player who massively increases their basketball IQ than who increase their shooting ability.”
Lowe and Falk thought one player might become a better shooter as he became a bigger shooter, getting “bumped off his spot” less. Body size is toward the malleable end of the spectrum.
Be there. The Sixers had a tumultuous ‘process’ of rebuilding – or at least that’s how it looked from the outside. But the inside was different.
“We got a lot of flack for not having the right ‘veteran leadership’ but you have to be on the ground and see it day to day get a sense of what the impact of that is.”
“As soon as Sam started there, there was a lot of noise. The media sort of latched on to this story and there were a lot of people with very strong opinions either way.”
The outside view can be helpful, like when we use base rates, but it can also be misinformed. During the Pittsburgh Pirates rebuild the media criticized signings because they didn’t see the same statistics. During much of Andre Agassi‘s life, the media criticized his actions because they didn’t know his life.
If you’re in those situations, says Falk, you need to find your own metrics of success.
“It’s easy to have your goals set from the outside and whether you succeed or not is determined by what the public reaction is and oftentimes that can be very misleading. There’s a lot of luck, maybe the public doesn’t right their expectations very quickly.”
“That’s easier said than done because all the stuff that goes on and when you’re with the Sixers and you’re not winning games how are you judging your own success?”
Metrics based on process remove the luck component that metrics based on outcomes include. This takes more work though. Outcomes; wins, losses, hits, runs, pounds, etc are easy to measure and compute in spreadsheets. But functions are not always faithful.
Alternative histories. Much of life is path dependent. We saw that with the series on Rory Sutherland who pointed out that much of the time we don’t decide what to eat but how. Once you make a choice you’re down a decision path and a host of other paths go untraveled. Falk recalled something similar.
“Who knows what happens if we win the game (in Boston), we break the losing streak, who knows if that changes things…all this stuff happens and that obviously changes the dynamics of an organization. That shifts some of the tasks you have to do. That shifts what you’re paying attention to at any given time.”
Incentives. Even though there were external pressures (from people who aren’t there).
“(Sam Hinkie) was incredible at having an internal consistency and not paying attention to whether or not something would cost him his job and paying attention to what was best for the franchise in the long run… a lot of times that’s how the worst decision in the NBA get made is people who are desperate to save their jobs. They’re acting correctly for their own incentives.”
Incentives – and not only money – matter. Wells Fargo, said Charlie Munger, had an incentive structure too aggressive and it led to people doing bad things. The best incentive systems align stakeholders and owners, principals and agents.
Numerators and denominators. Certain teams have good advanced metrics but aren’t winning games. Their process numbers look good but their outcome numbers look bad. Rather than always trumping outcomes with process, the curious will dig deeper.
“When you see a terrible defensive team with the right shot profile you wonder, are the shots so easy the opponents don’t even have to work for them?”
A team that allows ten shots at the rim where eight go in will have a better “Shots at Rim” statistic than a team that allows twenty shots at the rim but only five go in. If you don’t have rim defenders then the rate that rim goes up. Analytics don’t solve problems like a silver bullet. Analytics lets you ask better questions.
Be different. In his resignation letter, Sam Hinkie pointed out the importance of being different. We covered that here too.
In Philidelphia, they recorded three-point rates in practice. “We did things that were thought of as crazy,” said Falk. If a player shot above a certain rate in practice they got a green light in games.
It’s hard to be different, Zach noted. “They’re (the Pelicans) trying lineups right out of the jump that maybe would have taken them twenty or twenty-five games to get to had they been healthy and I think that might have been good for them.” Too often teams have an incentive structure that rewards tradition over interestingness. As Sutherland put it:
“One of the reasons stupid, or pig headed people do well, and when they do well they do really well, is because they are ignoring all the category norms everybody else thinks are important and they’re emphasizing something completely different.”
Using data. Falk created Cleaning the Glass and said:
“(About the site) I focused a lot on making the stats understandable. The ranks and percentiles are the key because there’s a lot of advanced stats that people don’t know what it means but if you know compared to other players at that position that tells you a lot more. One of the other things I did was broke things down by context.”
In Big Data Baseball we hear the story of the Pittsburgh Pirates staff getting buy-in from the players. One thing that led to this was understandable statistics. For the Pirates, heat maps were the key.
Lowe and Falk talk about a lot more; optionality, redundancy, luck, small samples, distributions, Bayesian updating, and weighted decision making.
Thanks for reading. If you want an email that includes all the best stuff I listened to each week it’s here.