Category: Uncategorized

Large N, small p (cancer, Netflix)

MIKE1 Comment

In addition to the first post, we can add two more ideas of small probability times a large number yielding a significant result.

In the first instance, our large N is t (time), and the small probability event is genetic mutations which lead to cancer. Jason Fung writes about mutations: “This small likelihood of success explains why cancer often takes decades to develop, and why cancer risk rises sharply in people over the age of forty-five.”

The second is an idea from Mario Cibelli about accumulating advantages. Cibelli told Patrick O’Shaughnessy that he visited a Netflix distribution center during their DVD heyday.

“I think what we saw essentially was an operation that was very, very hard to replicate. They had years and years of finding and bumping into bottlenecks and eliminating them, and getting more and more and more efficient. That would range from how labor was used, the lack of storage of DVDs. They actually didn’t store them anywhere, they always remained on the desks. The manager explained to us how the DVDs were always looking for a home. They weren’t trying to find the DVD that the home wanted, they would have the DVD in hand and say, “Hey, which home wants this?” To a bunch of machines that they bought that sorted the material that didn’t work, that destroyed a number of DVDs, and that they had to customize.”

Each obstacle was small, take many small improvements and you’ve got a business. Netflix’s small p large N effort was how they won the TiVo Race.

3 prompts to being Bayesian

MIKELeave a comment

How much should new information matter? Or, is this time different? Because sometimes it’s not.

“Merlin (Heidemanns) said that essentially the polls gain more weight. It’s not that we construct a model and weight the polls. We don’t take a weighted average of the polls, we estimate latent parameters and the polls are data. That said, you can roughly approximate the estimate as a weighted average.”

Andrew Gelman

According to Gelman, priors like “the economy, stupid” never exit the model. According to Nate Silver, the final poll removes all prior data.

In college sports priors matter more than Gelman allows for politics. Nearly two-thirds of college basketball teams who start a season ranked in the top-25, finish in the top-25. College football is the same.

How much new information should matter is a tricky question, but it’s helpful and why the Wharton Moneyball co-hosts encourage each other to become more Bayesian.

At the start of a football season we can guess (or hope) on a team’s chances. With more information, each play, game, and season, we update our idea. Eventually our guess at the start of the year gives way to the information of the year.

It’s hard to do though because we don’t know is this time different. Most of the time it’s not different enough, and base rates work best. But there are three general frameworks which might help us become more Bayesian.

First is to ask a Marc Andreessen like question: is software eating the world? How has the system changed and what does that mean? From FAANG to Testa joining the S&P, it seems like a systemic shift toward technology. Ditto for passing in football.

Second is to ask the Michael Mauboussin question: how much of this was luck? There’s a lot more luck in a single football play than an entire football game. It’s always a mix of skill and luck, but in what ratio?

Third is to consider our identity: am I attached to a position for unacknowledged reasons? This category includes biases like sunk cost and personal influences like ego or status.

The 2021 vaccine rollout is a good instance of practicing Bayesianism. Start with the base rate for vaccines. Watch for evidence. Adjust accordingly.

Every Business is Two Businesses

MIKELeave a comment

From Bill D’Alessandro on Circle of Competence #149.

The latest ILTB podcast here, provides examples specifically around media and content companies. Dave Portnoy being the starkest example.

And regarding franchising on Business Breakdowns. Along with never get in a land war in Asia, never take your burrito to Latin America— but your processes will probably do fine.

How jokes, and all things, work

MIKELeave a comment

Here’s Jerry Seinfeld telling Tim Ferriss about an idea he’s got. It’s still early. We don’t know yet if it’s a joke. Seinfeld said, “I don’t know what to do with that”

“When you’re on a cell phone call and the call drops, and then you reconnect with the person, they’ll go, “I don’t know what happened there.” As if anyone is expecting them to know anything about the incredibly complex technology of the cell phone, they offer this little, I don’t know if it’s an excuse or an apology. They go, “I don’t know what happened there.”

After Seinfeld has an idea he writes it down (there’s a lot of good writing and creative tips in this episode) and he works at it. Seinfeld explores the idea like my mother-in-law explores the home goods stores. Is this a good decoration? Does this match what else I have?

Seinfeld writes on yellow legal pads until a joke is pleasing to the ear. Then, it’s time to see how it works. And to remember, nothing is above the laugh.

At a comedy club the joke thrives, it dies, or it suffers enough damage to limp home and recover to emerge stronger and better prepared the next time. The comedy club is feedback.

“That’s the paradise of stand-up comedy. You don’t have to ask anyone anything. Stand-up comics receive a score on what they’re doing more often and more critically than any other human on Earth.”

Jerry Seinfeld

All things work like this. From idea to iteration to feedback in the market. Stand-up from Seinfeld is the cleanest version of this. Jerry’s method is the IKEA instruction of comedy, down to the simple paper it’s printed on. A comedian can have a joke in the morning, work it over over lunch, and deliver it after dinner.

All creations follow this process, but comedy is the gold standard because it’s clear and clean and quick. Write a newsletter (neé, blog) and the feedback is slow. Create a product and after development, marketing, and distribution you might know if people like it.

Poker’s appeal is principally the same. It’s cause-and-effect world. It’s easy to see. We like that. Comedy too.

Life is messy. But this helps. Keep in mind that the same process underlies everything creative: idea, iterate, feedback. The loop may be longer, but the process is the same.

How to complete next year’s NCAA Men’s Basketball bracket

MIKE1 Comment

A quarantine activity in my sister-in-law’s family is buying stocks. Stonks. Gamestonks? Each week, each kid, gets five dollars. They invest in whatever they want.

The best returns, to date, are from my niece. She’s 8. Her name begins with ‘T’ so her stock choices begin with ‘T’. She owns Tesla.

My three nieces and nephews aren’t competing but they do demonstrate that in small groups the best way to outperform others is ‘be chalky.’ Picking favorites is called chalky because in the days of horse racing, tracks wrote the lines on chalkboards with chalk. Even then, people liked betting favorites, so those odds were updated more often and had fresher chalk marks. Hence, ‘betting chalk’.

The same structure works for winning in any small group. To outperform, bet chalk. **However in a large group, choose variance.** Be different, and be right. We know that something will happen. We just don’t know which something.

One way to think through this approach is to consider the sum of the NCAAM final four team’s ranks. This question was posed on Wharton Moneyball and we have an answer: 11. On average, two number one seeds make it to the final four each year. Only in 1993 did all one seeds make it to the FF.

Yeah, but Covid!

That’s what I thought too. When Cade Massey proposed that it might be a more variable season I thought, base rates be dammed it’s going over. But that’s probably wrong (Narrator: It was not).

What I missed what something Daniel Kahneman wrote about in TFaS: substitution.

Rather than answer the question: Will the sum of the ranks of the final four teams be larger than average this year? I substituted the question: Will there be more variance this year?

What I missed was the idea behind hurdle technologies. In food preservation there’s not just one way that keeps food safe to eat, but a bunch. Food might be too acidic and be cooled and be sealed. It’s the combination of things, a series of obstacles, which limits bacteria.

That same idea applies in a bracket. Oral Roberts (#15) beat Ohio State (#2) and Florida (#7) but had to face Arkansas (#3) too, who won.

Ultimately the final four seeds totaled 15 (1,1,2,11). My direction was right. My reasoning was wrong.

Danny Meyer is an Alchemist (and you can be too)

MIKE1 Comment

“A question that stymied me for years and years, a question I got almost anywhere I went, it seemed like every organization asked, ‘How do you manage to hire so many awesome people?’.

“I said to look for fifty-one percenters, people who are emotionally wired to be happier themselves when they delivery happiness to you.”

Danny Meyer, ILTB

The central idea to Alchemy is to optimize important but overlooked things, with especially large returns from inexpensive yet important finds. How to find these things? Numbers provide a good clue.

When things are easy to measure, they are numerate. These numerate items are easy to discuss, to compare, to enter into spreadsheets then sum, average, and compare again.

Danny Meyer has succeeded (in part) because he competes in new areas. In the beginning of the podcast he tells O’Shaughnessy about competing on food and wine and ambiance and all that, but that’s what everyone does. It’s hard to have THE BEST food when everyone is trying to have that.

THE BEST food has convention. It has history. There are norms. There is price. Having THE BEST food in New York City is like being the best investor in New York. Good luck.

However, being the best at something slightly different is quite a bit easier. There’s a lot more area and a lot less competition if you do things off the beaten path.

Meyer found this in hospitality. Listening, we don’t get the impression that the tag wags the dog, but it’s got to be part of the reason Meyer is around, and talking on the Invest Like the Best Podcast.

Often Alchemy is using (free) psychology rather than (costly) structure. Better service rather than better linens.

One story Meyer tells in the book is about ‘the medicine cabinet’. One establishment was having the normal rumble of friction getting its legs under it and when patrons had a bad time, Meyer and staff offered a glass of dessert wine to soothe their pain.

Not only was the wine complementary, but it was special. At the time, dessert wines were novel so it was a special treat. The kicker was that they were the cheapest wines Meyer stocked.

Alchemy is like improving weaknesses, there is a lot of return for the initial effort, often much-more than optimizing factor f for the tenth time.

Danny Meyer is an alchemist. From the people he hires to the businesses he starts.

You are an alchemist. Find something that’s important but not measured, and deliver that.

2021 Predictions

MIKE4 Comments
My Superforecasting notes.

So, this probably should have been written in January. Writing it in March means it should be more accurate. Less time, less variance (see also: Something is always happening)

One of the lessons from thinking like Tyler Cowen was to see the world as it is, not as we’d like it to be. Making accurate predictions is one way to approach that concept. One of the lessons from Phillip Tetlock’s Superforecasting was that improving predictions is possible.

Tetlock gives 10 commandments for better forecasting, one of which is to practice forecasting. Here are the prediction, if you want an overview of Tetlock’s book see the post: Is Bill Simmons a Superforecaster?

Hurricanes. NOAA “An average season has 12 named storms, six hurricanes, and three major hurricanes.”

Will there be more than 12 named storms? Yes, 90%.

Will there be more than 30 named storms (the 2020 record)? Yes, 25%

Will there be 3 or more major hurricanes (top winds of 111+mph)? Yes, 60%

Will I lose power at my home in Central Florida for more than 3 days? Yes, 10%.

Blog.

Will this blog have more than 41,000 views in 2021 (41k is the 2020 number)? Yes, 15%

Will this blog have more than 800 posts by year end? Yes, 30%

Finance

BTC will top 75,000 at any point in the year? Yes, 10%

BTC will be under 30,000 at any point in the year (started 2020 at this point)? Yes, 20%

ETH will top 5,000 at any point in the year? Yes 5%

ETH will be under 130 at any point in the year (started 2020 at this point)? Yes, 20%.

BRKB will top 275 at any point in the year? Yes 10%

BRKB will be under 234 at any point in the year (started 2020 at this point)? Yes, 30%

Economic Recovery These will be graded per Bill McBride’s numbers on Calculate Risk

Any single day of the last week of the year will top 2M travelers (2019 was 2.0-2.5M)? Yes, 75%

Open Table reservations will be down less than 10% YOY? Yes, 75%

Open Table reservations will be positive YOY? Yes, 20%

Any movie earns more than 250M on opening weekend during the year (Dark Knight in 2019)? Yes, 5%

Hotel occupancy tops 60% (graph)? Yes 80%

Hotel occupancy tops 70%? Yes, 60%

Sports (I needed two more)

As of year-end, Tom Brady averages +270 ypg? Yes, 30% (This is an ongoing thing)

The Lakers are NBA champions? Yes, 25%

The Restricted Actions Section

MIKE8 Comments

Shane Parrish on Capital Allocators:

“Even during the pandemic, there are tons of public health guidelines out there where people are telling you what to think, how to think. You need to filter that and digest it. You can’t just rely on it. They came out and said that masks don’t matter and then said masks do matter—well there was no downside to wearing a mask. You might look like an idiot in the short-term but there’s no downside to it.”

Shane Parrish

One theme in Shane’s great conversation with Ted Seides was how much the cost of looking dumb restricts possible actions. About his podcast Shane said he’s just an idiot with a microphone.

But restricted compared to what? An idiot how?

Restrictions differ by scale and is much like the old political joke: I’m a libertarian at the national level but to my dog I’m a Marxist.

This idea surfaced in Bill Brewster’s podcast with Dan McMurtrie.

“I was somewhat bold enough to call out a transaction that some people have been burned on. But when I started to get inbounds from real mutual funds and managers and as I listened to why people weren’t buying I was like ‘Oh, I’m gonna fucking win on this because I don’t have the constraints.”

Dan is super good in the interview and together they address the four levels of constraints.

  • Macro culture (society). For instance, it’s taboo to talk about sex, or at least the dating market.
  • Micro culture (office). In another Seides podcast, he spoke with Ben Reiter about culture.
  • Position (job). Certain institutions have mandates about size, moral, or industry situations. ESG is a literal example.
  • Psychology (self). In the podcast Dan and Bill joke about investors who say “See’s Candy is my fav WEB investment”. That’s a psychology restriction.

So what? Why do constraints matter? Because they limit what a person can do. Dan again:

“I never want to compete against Stan Druckenmiller in timing the market. I never want to compete against David Einhorn in valuing a company. I never want to compete against Dan Loeb in writing an aggressive letter to a board. Where I might compete is where the environmental factors means that fighter is not able to perform at their best.”

It’s not more options that are better, but different options. Having good ideas (‘Go’) that also look good (‘Show’) is twice as difficult as just having a good idea. Or, can you look a bit like an idiot?

Correlated Decisions

MIKELeave a comment

“We use quantitative methods to put together diversified portfolios that don’t blow up over time. We have technical guys who are very sophisticated, one guy was the MIT chess champion. We need these guys to balance our portfolios, but they’re not picking stocks. I pick those guys because they have no idea how to pick stocks and I don’t want to know what they think about picking stocks, that’s for our researchers.”

– Joel Greenblatt, Capital Allocators.

The JTBD of diversity within an organization is uncorrelated decisions. If we’re all thinking the same way, the expression goes, nobody is thinking.

Early in the episode with Ted Seides, Greenblatt cautioned that there’s always more correlation in a portfolio than someone expects. This is consistent with the idea that most financial issues are liquidity issues.

If there are 3 ways to spend your day then we should be wary about the feed, the search, and the trends. For instance, the September 1 – December 1 BTC search trends. When Coinbase emailed me “Why Bitcoin is in the News.” I thought, this might be a good time to sell.

Google Trends for “bitcoin”.

Trends, searches, and feeds aren’t bad, but they are correlated.

The best solution to uncorrelated decision making is to use a bit of decision making advice from Rory Sutherland. When we select one at a time, we choose the average item. When we choose multiple things at a time, we choose a variety.

Rather than consider a best option then, we can consider a basket of options. With finite resources it’s hard (impossible?) not to prioritize but it does lead to new ideas.

It’s neat to hear that Greenblatt’s operation is like a newsroom: editorial and news, research and technical. Division is balance. The worst outcomes aren’t when something goes against us but when everything does.

How we decide

MIKE3 Comments
Our first boil, 2019

Moving to Florida in 2018 has been mostly positive. Y’all and ma’am are great and I use them each day. The weather is wonderful. We also make a boil.

A low-country boil (minus the pesky crawfish) is a holiday and weekend staple. Boil water with seasoning; add potatoes, corn, onions, sausage, shrimp, enjoy. The ratio of work to taste is very low. It’s a good deal.

But making it the first time was hard. We had friends over and I didn’t want to be the person who gave everyone food poisoning or served potatoes that tasted like dirt. The whole meal went great, if not a touch spicy, and each subsequent preparation has been slightly easier even for serving a crowd.

There’s a lot of other options we could do as well. We made a lot of chili when we lived in Ohio. We could order out, make sausage stuffed potatoes, or any number of things. But we don’t. We make a boil.

A boil is familiar. It’s easy. The opportunity cost is opaque. Will <other meal> be better? We don’t know. Let’s make a boil.

A lot of decisions are like this. Opportunity costs are hard to quantify.

During the late teens one bit of regular startup advice was that a product had to be 10x better than the existing option. While JTBD offers a slightly different approach, the idea is a good one. People do switch from one thing to another all the time but it’s often because the decision to do so is easy.

Making a boil in Ohio: hard.

Making a boil in Florida: easy.

Organizations then can consider how to dial the friction up or down. To keep serving people, make it easy for them to stay. To serve new people, make it easy for them to switch.