A feat few remember is that Sabean turned around turned around the 1996 94 loss team to a 90 win team in 1997, which won the NL West, but then they unfortunately lost to the wild card Marlins in the LDS. So while he got a good pick in his first draft in 1997, once you start winning, you get a lot of picks with poor odds of finding a good player, which continued until the Giants started losing in 2005. My contention over the years, because of my first draft study, was that it was very hard for a team to rebuild via the draft, because the odds were against it.
Sidenote: I've decided to create a separate blog where I'll be posting my R baseball analyses (when I'm able, still learning, part of my New Year's Resolution to get proficient in using R), probably starting with my draft database collected from baseball-reference.com, a great resource for anyone who wants to analyze baseball.
ogc thoughts
1997 Draft
This was a relatively successful draft, in that the signed players selected produced 13.8 bWAR over their careers. Unfortunately, most of that were produced for other teams, as Grilli and Linebrink, the two headliners (it's all relative), were traded. The probabilities added up to .3557 and based on the draft odds over the 50 year history of the draft if someone randomly selected from each population, there was 67.42% chance of ending up with no good player selected out of the seven picks.
1998 Draft
The Giants got a lot of extra picks in this draft, for a total of 11 picks among the first 200, as they lost a number of free agents. It was mostly for naught, only 3.0 bWAR generated. However, it is a well loved draft, as the Giants selected Ryan Vogelsong with the 158th pick overall, in the 5th round. As nicely as he pitched for us, he was not so good before or after, resulting in only 1.5 bWAR as his career total. But he produced 4.6 bWAR in 2011-12, and then another 0.9 in 2014, so he was very useful to the Giants over that period (he was injured in 2013, and had a poor season). Makes me wonder if I need to change the definition (not for my hand tallied analysis, I would try to use R to do it) to be, say, 1 to 8.9 bWAR is useful, and 9.0 to 17.9 to be average.
It maybe could have been much better. Per a source who said he had inside information, the front office (especially Dick Tidrow) wanted to draft CC Sabathia, but the ownership, in a pattern that continued throughout Magowan's tenure as managing partner (like punting picks and skipping on trying to sign Vlad), was unwilling to meet his bonus demands, so they instead drafted Tony Torcato.
It's been well reported that minority owners placed a lot of constraints on the Giants spending, as they were unwilling to answer capital calls to invest more into the team. That's why owners like Johnson, Burns and Neukom bought out those minority owners (Neukom sold out to Johnson when he was pushed out as managing partner), building up their shares over the years. But those constraints held back the Giants until the 2010's. Particularly in the early years after Magowan took over, the scouting department was decimated, and probably did not start recovering until Sabean took over.
The probabilities for this draft added up to 0.4886, making it a cumulative two year total of .8443. However, although there were so many higher picks due to the supplemental picks awarded for free agents (3 pick in Top 30, 5 in Top 41, 7 in Top 72), the probability was still a high 60.42% that the Giants would end up with nothing.
1999 Draft
Another bad draft but the first one that had a lot of potential, with Kurt Ainsworth and Jerome Williams drafted 19th and 25th overall. There were 8 picks with the good probabilities adding up to 0.2766 (three year total of 1.1207, meaning that on average, the Giants should have expected to find at least one good player). Still, it was a high probability that none were found, with 75.36% chance of no good player found. And over three seasons, 30.69% of the time the team would not have found a good player.
2000 Draft
A really dry year for production (only -0.3 bWAR) but notable for the drafting of Boof Bonser and Lance Niekro, out of the six picks. The probabilities only added up to 0.2286 (four year total of 1.3493) and 78.91% chance of not finding a good player (four year odds are 24.22% of not finding a good player).
2001 Draft
A return to okay draft production as they selected Brad Hennessey and Noah Lowry with their #21 overall and #30 overall picks, resulting in 9.8 bWAR. Then following them are two well known prospects, Todd Linden and Jesse Foppert.
If you'll recall from my first draft article, even if a team gets 1.0 bWAR from a draft, that should cover the entire cost of the bonuses from that draft. So the dry years like 2000 are covered by the better years like 1997 and 2001 (or even 1998-1999, where there was 5.6 bWAR produced overall). The probabilities added up to 0.3326 (five year total of 1.6819) and there was 70.91% odds of not finding a good player in this draft (17.18% of not finding a good player).
1997 to 2001 Draft
The first five drafts of Sabean's reign as GM was not what one would hope for in keeping a winning team winning. There were no good players found, and the closest was Lowry, with 10.0 bWAR (shows how good his career was, before it was cut short by his wrist surgery for thoracic outlet syndrome, which he sued the Giants for botching the surgery - I believe he lost that suit), just beating out Scott Linebrink, with 9.3 bWAR. Jason Grilli was third 4.6 bWAR, Jerome Williams next with 3.6 bWAR. Brad Hennessey rounds out of the Top 5 with 2.5 bWAR.
Still, all together, they added up to 28.9 bWAR, or 5.78 bWAR average per draft, which is still a great return on the bonuses paid out. In today's baseball economy, that works out to almost $70M worth of WAR, when MLB teams pay out anywhere from $5M to $16M in the 2019 draft bonus pool, based on slots. And usually pay out a bit more, as a team can go up to 5% above slot overall, plus can pay out $150K to any draftee after Round 10 without adding to the bonus pool for penalty purposes (and teams often underpay in Rounds 1 to 10, so that they can pay more in later rounds).
Low Odds Can Add Up Fast, But Not Fast Enough to Stay Competitive
What I'm surprised is that the probabilities, although low, can add up to an expected good player pretty fast. After five drafts, with only the one good pick in 1997, but a lot of supplemental picks, the Giants under Sabean were already up to 1.68 expected good players, with only 17% odds of ending up with nothing, meaning he was far below average at this point and . Of course, the rumored Sabathia pick would have fixed that, but I have not heard any corroboration on that selection other than the one source I had.
In any case, as surprised as I was, the discovery rate of good players, as shown by this data, is not that good. Empirically, I had looked at the discovery rate of consistent playoff contenders, and saw that the rate of discovery was very low, but still, this analysis shows that, like water on a rock, it adds up, enough that a team should be finding a good player via the draft, even while winning, every four to five seasons, depending on how many picks they end up selecting.
Of course, finding one good player every five years is still not a good replacement rate for a competitive team. St. Louis, a 91 win team in 2019 had 5 players that had 2.5+ bWAR (four with between 3.8 and 6.0 bWAR) and another 7 players that were average (between 1.5 and 2.5 bWAR), plus another 11 players with between 0.5 and 1.5 bWAR. Most of those five good players should be becoming free agents or reached their peak and started their decline phase in the next five years, with only new good one expected to be found from the draft to replace them.
Finding Great Players
Given this finding about the strong likelihood of finding a good player or better over a five year period, it got me curious about what the odds are for finding great players. So I did the same as above, but for odds of finding a great player:
- 1997: 0.2116 expected great player; 80.0% probability not finding a great player
- 1998: 0.1776; 83.6%
- 1999: 0.0915; 91.2%
- 2000: 0.0762; 92.6%
- 2001: 0.1212; 88.4%
With no great player found in this period (let alone good), obviously the Giants were behind the expected value, but a random team with similar draft pick selections would have ended up with nothing roughly half the time. So Sabean looks okay from the view of looking for great players, as he hasn't found one yet but he was just as likely to find one as to not find one. As we'll see, his luck changed drastically with the next draft.
Next analysis will cover the Dynasty draft years, from 2002 to 2009, from Cain to Belt, as the Giants continued to win, but eventually started losing.
I figured out a way to work backwards to get a ballpark number for the probabilities, using binomial probability calculations. Using the overall probability of not finding a good or great player, I could take the Nth root (where N is the number of picks) of that to get an equivalent same probability of not finding a player, which then gives me the equivalent same probability of each pick finding such a player.
ReplyDeleteFor finding a good player, the equivalent was 4.308%. That results in 17.2% chance of not finding a good player, 30.9% chance of finding one good player, 27.2% chance of finding two good players, 15.5% chance of finding three good players, and 9.2% of finding 4 or more.
Thus, even with that low rate of finding a good player, there is a pretty good chance (over 50%), randomly, to end up with two or more good players. Which makes Sabean's first five drafts utter failures.
The only saving grace is that it is very likely that Giants ownership was being cheap (until probably when Neukom took over), as evidenced by the punted draft picks, which could have forced Sabean to select players who are willing to sign, versus whether they thought they could succeed in developing.
For finding a great player, the equivalent was 1.726%. Again, as noted above, that ends up with 49.8% chance of not finding a good player. Then there is a 35.0% chance of finding one great player, 12.0% chance of finding two great players, and 3.2% chance of finding three or more, if viewed as random chance.
Together, these odds show that the Giants lack of success, as well as the Dodger's and Cards recent success (or the Giants success during the dynasty building period), could just be the result of random luck. We'll see what happens next for these two NL competitors.