• The Giambi Odds, Part II

    Posted by on February 27th, 2008 · Comments (12)

    Four days ago, the question “How many 37-year olds have posted great offensive seasons – even in part-time duty – for the Yankees?” was pondered here – which led to the conclusion that “…history tells us that we should not expect anything positive from Jason Giambi this season. And, should he actually provide some decent offense, it would be defying the odds – big time.”

    In response to that item, some asked to see the data for a larger population of 37-year olds – and not just Yankees – to see what those results would show. So, since you asked, click here to see the stats (via the Complete Baseball Encyclopedia).

    For the cut, we used 1973 as a starting season – as that’s when the D.H. came into play. And, we used 251 PA as a minimum to allow for part-timers.

    As you can see, there were 149 such “seasons” found. And, in 74% of those seasons, the 37-year old batter was not great. And, here, we’re turning a blind eye to batters like Barry Bonds, Rafael Palmeiro and Ken Caminiti who posted good numbers at age 37 via some help. Further, in 49% of those seasons, the 37-year old posted a bad offensive season.

    Using this bigger sample supports the suggestion made on Giambi four days ago. History here tells us that the odds are not in his favor. And, we should not expect anything hugely positive from Jason Giambi this season. Further, it’s just about a 50-50 chance that he should provide some decent offense at all.

    Comments on The Giambi Odds, Part II

    1. keithny
      February 27th, 2008 | 11:29 am

      Interesting, but we can always hope. A 2006 Giambi probably makes the Yankees the best team in Baseball.

    2. williamnyy
      February 27th, 2008 | 11:38 am

      This follow-up post contains the rationale that was completely missing from the original. In the latter, a rather low absolute BEST case was stated, whereas in this piece, more sensible words like odds and expectations were used. Should we expect Giambi to have a great season? No. But, does he have a chance to better the 20/71 benchmarks in the previous post? Absolutely. The biggest question for Jason is how healthy can he remain. If he gets his ABs, I think the odds are pretty good that he will be very productive from an offensive standpoint.

    3. Raf
      February 27th, 2008 | 11:41 am

      Thanks for putting out the analysis.

      One thing I noticed is that there are a lot more players playing “late” in this decade than the decades before.

      Given BR’s similarity scores, at the high end, he could be Andres Galarraga, at the low end, he could be Jim Edmonds.

    4. Lee Sinins
      February 27th, 2008 | 11:46 am

      There is a flaw to your methodology here.

      A 50-50 chance that a player provides some decent offense at all isn’t just a description of 37 year olds. It is a description of major leaguers.

      Half of them are above average. Half are below average.

      I just did some calculations with the raw data that shows there is absolutely no correlation between age and RCAA.

      There is literally no age in which the average RCAA of players that age is greater than 1.

      With the exception of real young players (16, 18, 20 & 21 year olds have an average of -2 and 17 & 19 year olds have an average of -3) and very old (47 year olds have an average of -4, but that’s an extremely tiny sample size of just 4 players), there is no age in which the average is worse than -1.

    5. williamnyy
      February 27th, 2008 | 11:58 am

      I am not sure how useful it is to compare individuals to the average of all players a certain age. For example, in determining the odds of Arod being above average this season, I’d be more interested in using accomplished 32-year olds as a benchmark as opposed to every 32-year old in the league. Similarly, I think Giambi needs to be compared against 37-year olds with similar accomplishments (and body/injury profile), not every 37-year old who played the game.

    6. Lee Sinins
      February 27th, 2008 | 12:03 pm

      William,

      We’re both really saying the same things.

    7. williamnyy
      February 27th, 2008 | 12:19 pm

      Yep…you are questioning the methodology and I am questioning the premise. I think this analysis fails on both counts.

    8. February 27th, 2008 | 1:11 pm

      “A 50-50 chance that a player provides some decent offense at all isn’t just a description of 37 year olds. It is a description of major leaguers.”

      Maybe I’m going about it the wrong way? But, here’s the point. If a 30-year old is coming off 4 straight years of 25+ RCAA, then I would think the odds of him posting another season, at age 31, of 25+ RCAA is much higher than 50-50. Whereas, a 37-year old coming off mixed seasons is less likely, meaning 50-50, of posting a high RCAA season. This make sense?

    9. williamnyy
      February 27th, 2008 | 1:27 pm

      I think that was Lee’s point…you can’t simply compare a player to everyone else who shares his age, which you did by searching for seasons of 37-year old players. Based on Lee’s data, if you insist on such a framework, every player will be a 50/50 proposition. While there may be a basis for your conclusion, the premise and methodology used doesn’t support it.

    10. alvarof
      February 27th, 2008 | 2:08 pm

      How about Giambi’s PECOTA for 2008? Anyone has it?

    11. Lee Sinins
      February 27th, 2008 | 2:13 pm

      William’s correct.

      The 31 year old is a good bet for 25+ RCAA because he’s had 4 consecutive years with 25+ RCAA. That’s the relevant consideration, not the fact that he’s 31.

      Only 9.9% of 31 year olds have 25+ RCAA. But, the track record of that particular player is what is relevant, not what percentage of 31 years have 25+ RCAA.

      In the case of Giambi, what’s relevant is his track record.

      In the past 3 years, he’s had 49, 46 and 4 RCAA. A bad figure would appear to be next in the sequence.

      On the other hand, I am not as pessimist about Giambi as most people. His 5 years prior to those 3 years, he had 100, 103, 75, 51, -4. A series of bad figures would appear to be next in that sequence, but he follow that with 49, 46, 4.

    12. Sherard
      February 27th, 2008 | 4:05 pm

      “Staff” said – “If a 30-year old is coming off 4 straight years of 25+ RCAA, then I would think the odds of him posting another season, at age 31, of 25+ RCAA is much higher than 50-50. Whereas, a 37-year old coming off mixed seasons is less likely, meaning 50-50, of posting a high RCAA season. This make sense?”
      ————————————————-
      Sure, that makes sense, but that’s NOT what you did. If you REALLY want to know what to expect, historically, from a 37 yo, then you do exactly that kind of analysis. Of those 47 seasons, break up the players careers into thirds, for instance, or quarters. Break them up into below average, average, and above average. Then get a baseline for their RCAA for the first 2/3 or 3/4 of their careers and then compare that to the RCAA in the latter part of their careers and their age 37 season.

      THAT would tell you what to expect of a player at age 37. I would assume that performance goes downhill compared to their career averages, but how much ? Saying “He won’t be great” is pretty damn lame.

      I would also look for any correlation that shows the performance dropoff decreasing through the years. I would easily bet that the dropoff for 37 year olds was much greater in 1973 than it is in 2008.

      Additionally, don’t necessarily restrict it to the DH, as it seems pretty clear that the Yankees intend to play Giambi at 1B as much as possible.

    Leave a reply

    You must be logged in to post a comment.