+-----+-----+-----+----+-------+----+----+-----+------+------+
|  R  | HR  | RBI | SB | AVG   | W  | SV |  K  | ERA  | WHIP |
+-----+-----+-----+----+-------+----+----+-----+------+------+
| 778 | 186 | 705 | 96 | 0.290 | 72 | 62 | 978 | 3.77 | 1.26 |
+-----+-----+-----+----+-------+----+----+-----+------+------+

+----------------------+----+----+-----+----+--------+
|                      |  R | HR | RBI | SB |  AVG   |
+----------------------+----+----+-----+----+--------+
| Roster Doctor Team   | 86 | 21 |  78 | 12 |  0.290 |
+----------------------+----+----+-----+----+--------+
| Yahoo 1st Place Team | 94 | 25 |  92 | 18 |  0.294 |
| Yahoo 2nd Place Team | 91 | 23 |  89 | 16 |  0.290 |
| Yahoo 3rd Place Team | 89 | 22 |  87 | 15 |  0.287 |
| Average Yahoo Team   | 81 | 21 |  79 | 11 |  0.285 |
+----------------------+----+----+-----+----+--------+

+----------------------+----+----+-----+------+-------+
|                      |  W | SV |  K  | ERA  |  WHIP |
+----------------------+----+----+-----+------+-------+
| Roster Doctor Team   | 10 |  9 | 140 | 3.77 |  1.26 |
+----------------------+----+----+-----+------+-------+
| Yahoo 1st Place Team | 10 | 16 | 126 | 3.27 |  1.20 |
| Yahoo 2nd Place Team |  9 | 13 | 120 | 3.43 |  1.23 |
| Yahoo 3rd Place Team |  9 | 12 | 116 | 3.54 |  1.24 |
+----------------------+----+----+-----+------+-------+ 

---

Posted by Paul Singman at 12:29am (0) Comments

Hit Rate Observer

Question: How long is a baseball season?

OK, OK—it’s six months. But…

How fast is a baseball season?

One of my recurring premises about fantasy is that baseball players are awful, awful investments. For one thing, they are human beings. That means that they have a will, which they can direct (or not) to outcomes worthy (or not) on a steady basis (or not). So the player you get in one year may have different priorities in the next. And there’s little or no buffering—a baseball player isn’t a group of people working in concert and tempering the fortunes of the larger entity. Milton Bradley comes with dice, but they’re not made by Hasbro.

Also, as human beings, baseball players’ work is affected by external circumstances—marriage, children, Playstation, Alyssa Milano. More to the point, your investment in a ballplayer can plunge to zero in a vast number of ways, most of which are untied to any previous warning sign. Companies do go bankrupt (and sometimes very quickly), but never in one day, and never because the guy in Big Toe crashed his forklift.

The crux is that baseball players decay very quickly. But compared to what? We need a benchmark.

Since we have already compared ballplayers to companies, and since companies are familiar investments, let’s use them. I have a database, from 1955, of all Fortune 500 companies—the “big leagues” of business. Let’s compare the decay rate of companies against that of hitters.

The Fortune 500 measures only gross revenues, not profits, so instead of trying to pick an offensive metric by which to rank hitters, we’ll go with straight PA. After all, a player with a high PA has some combination of good focus, good durability, and little competition. In whose productivity could we have greater confidence?

For each group (companies and baseball players), we’ll track the survival rate of the top 200 members. For hitters, that’s a threshold of 400 PA—a good demarcation for fantasy value. We will follow the performance of the groups from 1995 to now, tabulating at each year the number of members that remain in the top 200.

Here is the chart for the Fortune 200:



The rate of attrition is far from devastating: Even 13 years on, about half of the Fortune 200 companies from 1995 are still in the top 200. Long term, the rate of decay in the Fortune 200 is about nine companies per year.

And the rate for hitters?



The long-term attrition rate is 16 batters per year—double the rate of the Fortune 200. The 2007 reunion of the fantasy class of 1995 has only 1/8 as many invitees as does the business reunion from the same year.

Let’s do the same thing for pitchers. Again, we will look at the top 200 members (a cut-off of 65-70 innings):



The long-term attrition rate for pitchers is 17 players per year, similar to that for hitters. However, the drop-off after the first year is steeper: The average number of pitchers remaining in the Top 200 after one season is 133—a loss of 33%. For hitters, the comparable number is 143 members. The number for companies: 180.

And that’s why fantasy baseball is so frustrating. We play a game where 1/3 of the most productive players this April are gone by the next—not necessarily out of baseball, but probably out of the class of rosterables. Ballplayers decay twice as quickly as do companies in the long term, and more than three times as quickly in the short term. Predicting a ballplayer’s output next year is like forecasting Exxon’s revenues in 2012.

As fantasy players, we can adapt to this finding in two ways. The first concerns how we run our leagues. We should try to mitigate (not eliminate) the frustrations built into the game. The best way is to appeal to owners’ competitive spirit.

One idea: Grant partial pay-outs—at least by 1st and 2nd half, and maybe even by month. Owners who lose a key player in April may still be able to regroup and recover some of their entry fee by August.

Also, permit departing owners to sell their teams to incoming owners. In other words, a new owner could pay $20 for the departing team’s roster, or start from scratch for free. That creates an incentive for fed-up owners to leave their teams in creditable shape. (And they might work so hard that they decide not to leave after all.)

Our second adaptation concerns projections. Clearly, fantasy players have an interest in projections. But we need greater humility—there are just so many obstacles to our foresight between now and September. One could argue that seasonal forecasts should be stated in units no smaller than 25 PA, and maybe as large as 100 PA.

Next week, we’re going to start rolling out weekly updated projections in HEATER. However, guided by the above sentiment, we’re going to tread more modestly. We’ll offer projections in two flavors:

• Point-in-time estimates of “true talent” (BA/OBP/SLG for hitters; K/9, BB/9, HR/9 for pitchers), not associated to any projected playing time. We think that people who are mulling trades are weighing numbers like these in their heads. They probably have some idea of anticipated playing time, but it’s a broad and vague range, and we doubt that we can come up with much better. What we can provide is the best estimate of the player’s talent as of this moment.



• Projected hitting and pitching lines for the coming week only. After all, if we can’t predict playing time for six days ahead, we shouldn’t be predicting it for six months ahead. Curiosities to some, these forecasts are useful for owners in weekly leagues, or people looking to replace a guy on the DL.

A final lesson? If you are anywhere near the title, don’t give up. Somewhere, somebody’s caller ID is reading A MILANO. With luck, it won’t be one of your guys.