Jeremy Lin is the talk of the NBA. Sportswriters everywhere are busy cranking out column inches on what people have called the ultimate Cinderella story: The emergence of an Asian-American Harvard graduate, seemingly from nowhere, as one of the NBA’s biggest stars.
On February 3, Jeremy Lin was the Knicks’ third-string point guard. Less than two weeks later, after leading the Knicks to six straight wins, including two consecutive games in which he scored the winning points, Sports Illustrated announced that they would put him on the cover of the February 20 issue with the caption, “Against All Odds.”
Yet one group wasn’t surprised by Lin’s success. A new breed of basketball statheads (the hoop equivalent of the SABRmetricians popularized in “Moneyball“) had predicted Lin’s success from the start.
Prior to the 2010 draft (where all 30 NBA teams passed on Lin) their analysis ranked Lin #10 out of all players, and #1 among undrafted players. This analysis is purely statistical; the models don’t consider height, vertical leap, foot speed, and perhaps most importantly, skin color. They simply look at statistical contributions made during basketball games.
Statistical analysis continued to rate Lin highly on his rookie season. He produced .157 wins per 48 minutes played, or more than 50% better than the average player, who produces .100 wins per 48 minutes played. (Incidentally, Carmelo Anthony produced .140 wins per 48 minutes played that season).
He also shone in the NBA’s Developmental League (a minor league of basketball), where he produced at a .211 clip.
In other words, when you looked at pure production, Lin was a top prospect. His rise only seems unlikely when you consider non-basketball factors, like his race or educational institution.
Trendy sports blog Deadspin tweaked the madness best, titling a February 7 blog post, “Asian Harvard Grad Somehow Succeeding In New York.”
It’s a funny one-liner, but it underscores a more serious issue.
Lin’s high school coach noted that his star player wasn’t recruited by any colleges, despite leading underdog Palo Alto High to the California state title. He also noted that the following year, a number of scouts came to games to watch another of his players who wasn’t as good, but was African-American.
Stereotyping has legitimate purposes. If you knew that Harvard University had produced twice as many presidents (8) as NBA players (4), you would be right to guess that any generic Harvard basketball player would be unlikely to make the NBA. But stereotyping only makes sense in the absence of better data.
In the case of Jeremy Lin, publicly available statistics proclaimed his value, but scouts preferred believing in stereotypes to trusting in data.
Sadly, this kind of bigotry isn’t limited to the world of sports. Even here in Silicon Valley, where we like to think of ourselves as a meritocracy, we practice a particularly pernicious form of stereotyping on a daily basis.
Investors love to talk about “pattern matching.” A common expression is, “I’ve seen this movie before.” There’s a reason why entrepreneurs constantly pitch themselves as “the AirBnB of ice skating” or “the iPhone of Valentine’s Day cards” (hmmm, that actually doesn’t sound so bad….).
This made sense in the absence of better data. When investors had to make decisions based on a PowerPoint deck and some rough prototypes, falling back on stereotypes was a good strategy. Indeed, I like to describe the default investing strategy of Silicon Valley as “invest in charismatic 20something Computer Science graduates from Stanford, MIT, and CMU (with Berkeley, UIUC, and Harvard as fallbacks), as long as they’re male and either Caucasian or Asian.”
In today’s world, with the ability to judge entrepreneurs based on a vast amount of publicly available data, ranging from social media to GitHub, with the ability to launch MVPs and generate tangible engagement and conversion statistics without raising money from investors, we now have the better data we need to make stereotyping AKA “pattern matching” AKA bigotry obsolete.
But old habits die hard. Just in the last few months, we saw a CNN special on black entrepreneurs in Silicon Valley. Whether or not you feel that CNN used ambush tactics to help stir up controversy, the fact is that African-Americans make up only 1% of venture-backed entrepreneurs nationwide. And just last month, Whitney Hess conducted an analysis of top venture capital firms that showed that the most gender-balanced firm was Kleiner Perkins at 23% female, while the majority of those firms had zero female investors.
Discussing such topics makes people in Silicon Valley uncomfortable. Few of us like to think of ourselves as racist or sexist. Yet I know of many entrepreneurs who feel that they are overlooked because of their skin color, gender, age or simply because they didn’t go to the right schools.
Jeremy Lin has been called the Asian Tim Tebow (or is it that Tim Tebow is the white Jeremy Lin?); we need to extend the lessons of Jeremy Lin beyond sports to the startup world. Decisions need to be based on performance on the field of play, not race, gender, age, or education.
And for those who are the first to recognize “pattern matching” for what it is, the rewards can be great. How many of those other 29 NBA teams could use Jeremy Lin on their team right now?
UPDATE: #Linsanity continues. Jeremy Lin had 10 points and 13 assists in 26 minutes of play during a blowout win over the Sacramento Kings. 7 wins in a row.
UPDATE: Looks like this post struck a nerve. Head on over to Hacker News and upvote, please.
UPDATE: The news gets worse for NBA GMs. Pre-draft tests show that Lin is actually faster than NBA MVP Derrick Rose, as well as #1 picks John Wall and Kyrie Irving. Incidentally, Lin’s Knicks beat Irving’s Cavs last night. Lin had 19 points, 13 assists, and perhaps most importantly, only one turnover.
15 thoughts on “Jeremy Lin, Women in VC, and the Bigotry of Pattern Matching”
Amazing post Chris. As the dad of a multicultural family (we are Korean-Ethiopian-Americans :>), I look forward to a day when data and results define our opportunities. It's coming.
"ambush tactics"? asking arrington a simple question during an scheduled on camera interview that his pr people set up isnt exactly an ambush.
Thanks! I've noticed a lot of reaction on Twitter from people with diverse cultural backgrounds.
Arrington characterized the interview as an "ambush". Whether it was or not is left to the judgment of the reader, and you're certainly welcome to point out facts that you think germane.
This sounds like it should be an Academy award winning movie starring Brad Pitt and Jonah Hill.
Chris — I agree that pattern matching can often produce counter-productive results. Unfortunately, it's hard-wired into out brain. It's how we deal with the vast amounts of data we input every second. It's how we know that something we've never seen before is a chair, or a car or a person. Our brain is *always* going to stereotype data based upon historical data — the challenge is recognizing that you're doing it and question whether it's valid.
Hope you're doing well, amigo!
I just love saying the following phrase: "Academy Award Nominee Jonah Hill." I hope they use that in the "21 Jump Street" ads.
Pattern matching and stereotyping are very important coping mechanisms…the point is that when there is enough data, they become counterproductive. I pattern match all the time; just a few days ago, I was talking with someone and uttered the words, "A bunch of MIT grads? Sounds like a winner to me!"
But the key is that in that case I had no other data on the startup. Were I to talk with them, I'd focus on more tangible metrics and evaluations.
I like your quote: Stereotyping only makes sense in the absence of better data.
It acknowledges that yes, we almost all rely on stereotypes from time to time.
It also points out there are definitely times to discard them.
I think this is a very realistic assimilation of the type of thinking advocated in Blink and what we grow up learning (i.e. more thinking is better).
Yes, Tebow is just a whiter, dumber, and more obnoxious NFL version of Lin.
Thought you might enjoy this Pinterest board…I do. The more women you invest with, the more you'll invest in. Pattern matching trains us to trust the familiar.
Thanks for the pinboard! I hope you don't mind that I shared it on Twitter as well!
As a Asian Woman entrepreneur, I've been through lots of stereotype. In Asia, woman's success are defined by playing as a good wife and lovely mother. Sometimes the success of woman will be viewed as the reason for the failure of family. It is really disappointing. Hope your article will break more walls and bring up more innovation to the world.
Terrific post and terrific blog, Chris. I'd love to subscribe my email or even RSS it but I can't see a way to do either. How can we best follow this blog? Please email to let me know and/or post a reply here. Thanks! Your new fan, Joy Chen
Chris – wonderful post. Thanks. Nikhil
Great post Chris. I'd love to hear your thoughts on what data angels / VCs should examine when considering an investment. Correlation Ventures recently raised a first round of funding and their decisions are based mostly on a large proprietary set of data they have collected about venture investments.
"Yet one group wasn't surprised by Lin's success. A new breed of basketball statheads (the hoop equivalent of the SABRmetricians popularized in "Moneyball") had predicted Lin's success from the start."
I think this statement greatly exaggerates the truth Chris. Yes, one statistical model predicted that Lin could turn out to be a solid NBA player, but in the entire universe of stat-heads only one, a FedEx delivery guy named Ed Weiland, actually went on the record saying Lin was one of the top point guards in the 2010 NBA Draft. The Wages of Wins Journal doesn't even meaningfully discuss Lin until after his breakout performances with the Knicks. Stat-heads simply did not predict that this guy would play this well, and are just as surprised as everyone else.