MASTER NOTES: Buy and hold

Many years ago, I was a much more active stock-market investor than I am now. I know, I know—the stock market is not analogous to most fantasy baseball formats. Time horizons are longer and share ownership is not exclusive.

Because I was still working and trying to build a retirement nest egg, I was willing to embrace more risk. You might say I was like a guy building a roster for a fantasy keeper or dynasty league. To find growth, I had assigned 15% of my investment portfolio to growth stocks—accepting the higher risk involved.

In those days, high-growth/high-risk meant Internet stocks. I put some money into a small company at around $24. I understood and believed in the company’s relatively simple business model, which was to compete in selling a niche product by accepting smaller (but still impressive) margins than the ridiculous markups at regular retail.

That model still holds, by the way. We see it in online sellers of such high-margin items as men’s razors and blades, women’s cosmetics, contact lenses, printer ink... you get the idea. I once read an article that said the then-new owners of the LA Dodgers were accepting lower margins than other teams to build better teams, which make more money.

Anyway, I cashed in my shares after about six months for a nice 70% gain. My family was growing, so I was reducing risk in my investments and reducing the “churn” of buying and selling in my account. I became a buy-and-hold guy.

You probably know where this is going. While the dot-com collapse saw hundreds of tech high flyers go bust, my former Internet stock rebounded. The company grew beyond its original niche—selling books online—and expanded into selling... well, everything online. Of course, the company was Amazon. And had I just hung on, each $1,000 of my investment would be worth ... go on, guess. I’ll tell you later.


The moral of the story is something that most honest stock-market analysts and brokers (yes, they exist) will tell you: Don’t try to time the market, especially on relatively short-run information. Buy good companies and hold them.

This all came to mind recently when I was looking at my Tout-AL team, which is languishing in the pitcher decimal categories. One problem is that I have been an active investor. I’ve been trying to time the market by churning my pitching staff based on various factors, including strength of opponent, recent high-pitch-count games, whether the pitcher was coming off a good start, a bad start, a two-start week, and astrological signs.

OK, I’m kidding about the last one, but I have tried most of the others, and it just hasn’t worked. The aim, of course, is to activate a pitcher who should do well and reserve one who shouldn’t. But my own experience is that I have been just as likely, or more likely, to miss good appearances while absorbing the punishment from bad appearances.

Curious about this, I wanted to check the validity of various “streaming” factors. I listed all the pitchers with 20+ starts. Then I let Excel randomly pick pitchers from tiers of perceived pre-season auction value:

  • From the $20-$30 tier, Jacob DeGrom
  • From the $10-$20 tier, Marcus Stroman
  • And from the $1-$9 tier, J.C. Ramirez

The first factor was strength of opposition, easily the most common streaming factor. Sit your pitchers against strong oppo, right? I divided all the teams into tiers based on team OPS+, an OPS metric adjusted for league and park:

  • The top tier (T1) was teams with OPS+ of 100 or higher (HOU, WAS, LA, TAM, DET, NYY, CLE, MIA, SEA)
  • The bottom tier (T3) was teams with OPS+ less than 90 (TOR, KC, PHI, PIT, SD, SF)
  • And the middle tier (T2) was teams with OPS+ of 90-99 (all other teams)

And here’s the thing: sitting against Tier-1 strong opponents and pitching against Tier 2 and 3 was counterproductive:

  • DeGrom’s ERA vs those poor offenses in T3 is a boon to a fantasy team’s stats, but his production and skills vs T1 teams were far better than against T2. If we were broadcast-booth ninnies, determined to create a narrative to “explain” these stats, we would probably say something like, “DeGrom is a real competitor—he really gets amped up against those top teams.”
  • Stroman hasn’t actually had any starts vs T3 teams, one of the joys of pitching in the AL East when the only T3 team in the division is yours. But his 2.50 ERA vs T1 teams is 1.25 better than against T2. Another amp-up, I guess.
  • Ramirez was best vs T2, but starting him vs. T3 teams was a disaster for fantasy owners, with an 11.25/1.47 line. New narrative, please!

The other widespread streaming play is to bench any pitchers going after his previous start had a high pitch count (PC). (For high PC, I used any start in which a pitcher threw 5+ pitches more than his overall median.)

This streaming tactic would have worked well with DeGrom, who had a 6.00/1.63 line in his four post-high-PC starts, compared with a 2.72/1.05 in games not after high-workload starts. The poor performance carried into strikeouts (down two k/9), walks (up two bb/9), and Cmd (1.8 vs. 3.8).

But it was a different story for Stroman and Ramirez. Stroman’s WHIP in post-high-PC starts was no different from other starts, but his ERA was a run-and-a-half lower, and his Ctl a walk-and-a-half better. Ramirez showed the same pattern: similar WHIPs but a much lower ERA in post-high-PC starts. But Ramirez’ skills were noticeably worse in those latter games, with Dom 2.7 lower, Ctl 2.5 higher, and Cmd an egregious 1.1.

Finally, various other tactics for streaming showed no patterns with these three pitchers. Sit a guy after a bad outing, start a guy after a good outing, start him when his astrological sign is rising in Cassiopeia with the moon ascendant, it doesn’t seem to matter.

And for those who always start pitchers going twice in a week, you might as well keep doing that. Without getting into details, skills for all three pitchers were almost exactly the same in two-start weeks as they were in the season as a whole. So you get more Ks. But the wins stayed random—percentages in two-start weeks were better for Ramirez, nearly the same for Stroman, and significantly worse for DeGrom.

Remember, of course, these are only three pitchers in around 20 starts and barely 500 batters faced. A small sample, in other words. But isn’t streaming pitchers, especially good ones, making plays based on very small samples? Benching a guy after a high-workload start? has run analysis showing that high-pitch-count starts do not predict poor follow-up performance (quite the opposite, in fact). Maybe there’s research work to be done here. Where’s that Costco bill again?

For right now, absent any proof to the contrary, I’m going to go with the Warren Buffett strategy: buy good pitchers and play them. Inherently, good pitchers will tend to pitch well, and should be started every time (barring actual news through due diligence—a sore arm, maybe short rest). Poor pitchers will tend to pitch poorly. There will be some occasions when those truisms will not bear out. But it seems like a mug’s game to try to “time” those occurrences, which seem pretty random and therefore unpredictable.

Oh, and that Amazon investment? As of this writing, each $1,000 of those Amazon shares would be worth a cool $480 grand. And I would be writing this on a Mediterranean beach, and not in a Waterloo, Ontario basement.

Click here to subscribe

  For more information about the terms used in this article, see our Glossary Primer.