Sep 27, 2012; Philadelphia, PA, USA; Philadelphia Phillies left fielder Darin Ruf (18) hits a three run double during the first inning against the Washington Nationals at Citizens Bank Park. Mandatory Credit: Howard Smith-USA TODAY Sports

Empirical Probability and Odds of Major League Success- Off the Radar

When looking at prospects in the minors, it seems that the hardest thing to project is whether a hitter will hit in the big leagues or not. There are more extreme cases like when scouts are looking at 16 year olds in the Dominican Republic and try to project whether or not they can hit at the big league level eventually (or even in the minor leagues). The more down to earth examples are guys who hit well in the minors, only to flounder in the Majors. One can find so many examples of this that my (ever revolving) philosophy has been to just get good and elite defensive prospects (value that is harder to measure than offense, but more projectable) and hope at least some of them hit, so you have some known value and have high ceilings. However, this doesn’t work as well with 1st baseman, and eventually, no matter how good you are at stealing bases or playing defense, each team needs someone to hit some homers. So while there has been some very good work done in the past on correlation between hitting statistics between the minors and majors, I wanted to do some of my own, this time, only using 1st baseman.

For this, I am using basic minor league statistics, that you can easily calculate using Baseball Reference. I use advanced statistics from FanGraphs, First Inning, and Minor League Central, but those statistics have not been available until recently, and for our purposes, we had to be able to go back to the 80s even in some cases. I looked at all the players that had at least 200 plate appearance as a first baseman in the Majors since 2000 and calculated their K%, BB%, K%-BB% (usually used as a pitcher statistic, but I wanted to see if it worked for hitters), ISO (I didn’t want to use things that were reliant on parks/levels etc., but I needed something beyond just K/BB for hitting, so I used ISO. As you will see, it worked out nicely), Plate Appearances, Round Drafted, and whether or not they were ever a Baseball America top 100 prospect (something else that makes Baseball Reference easy and enjoyable to use).

Plate appearances may have a survivor bias, especially for guys like Daric Barton and Dan Johnson (especially for a guy like Chris Richard, who became a full time AAA player after his MLB career), who spent significant time in AAA after reaching the Majors. Obviously the undrafted players are a mixed bunch (both foreign signings and players just not drafted), so I won’t measure them when looking at drafted players. If the player ever appeared on a Baseball America top 30 list, I noted that by putting yes. For round drafted, I only used the round they were drafted in when they signed.

I took Erubiel Durazo (Mexican League), Hee Seop Choi (KBO), Daryle Ward (Independent League), Jose Offerman (Mexican League and Independent League), Randall Simon (Independent, Mexican, NPB), Fernando Seguignol (Independent, NPB, Mexican) and Calvin Pickering (Independent League) out because of statistics in other leagues that affected their overall minor league statistics. I wasn’t sure what to do with Bobby Bonilla, as his 2000-2001 numbers were not reflective of his career numbers, so I took him out. John Olerud never played in the minors except for a brief rehab assignment when he was with the Red Sox. I didn’t include his rehab numbers, but kept him in the spreadsheet, leaving his numbers blank. Here, we are looking at the percentage of success each statistic (over each individual’s total minor league career) had.

League average wRC + for first baseman in 2012 was 107. So I am using that as the barometer of “success” and “failure”:

BA top 100 prospects: There were 34 failures, and 39 successes (~47%). Of the 84 that were not ever top 100 Baseball America Prospects, 24 were successes (~29%).

The median MiLB ISO for 1st baseman was .185. 31 of the 76 players over the median were successes (~41%). 25 of the 80 at the median or below were successes (~31%).

44 of the 1st basemen were drafted in the first round. 20 of them were successes (~45 %). 9 were drafted in the 2nd round, and 4 of them were successes (~44 %). 11 were drafted in the 3rd round and 4 of them were successes (~36 %). 4 were drafted in the 4th round and just 1 was a success (25%). 4 of the 12 drafted in the 5th round were successes (~33 %). 6 were drafted in the 6th round and just 1 was a success (~17 %). 8 were drafted in the 7th round and 1 was a success (~13%). The 8th round broke the trend, as 3 of the 5 were successes (60%). 1 of the 4 drafted in the 9th round were a success (25%). 29 were drafted in the 10th through 19th rounds and 10 were successes (~34%). 3 of the 10 drafted in the 20th round or after were successes (~30 %).

Median minor league plate appearances was about 2480. 7 of the 11 that had less than 1000 minor league plate appearances were successes (~64%). 43 players had between 1000 PAs and 2000 PAs in the minors, and 22 were successful (51%). 26 players had between 2000 PAs and 2479 PAs in the minors, 11 were successful (~42 %). 27 had 2480-3000 MiLB PAs (~42 %), and 12 were successful. 35 had between 3000 and 4000 MiLB PAs, and 10 were successes (~29 %). 15 had more than 4000 MiLB plate appearances, and just 1 was a success.

Of the 11 that struck out 10% of the time or less, 3 were successful (~27 %). Of the 49 that were between 10.1-14.9 K%, 16 were successful (~ 33 %). Of the 65 that were between 15-19.9%, 26 were successful (40 %). The 31 that struck out more than 20% had 12 successes (~39%).

For walks, the 7 out of 10 players (70%) with walk rates over 15% were successful. Those that walked over 10% of the time but less than 15% were successful 31 out of 73 (~43 %) times. Those that walked between 8 (usually about league average in the Majors) and 9.9% of the time were successful 16 out of 43 times (~37%). The 29 players with below 8% walk rates had just 7 successes (~24%).

16 players had more walks than strikeouts, and 7 of them were successes in the Majors (~44 %). The median was about 6 K%-BB %. Those over the median (that is, worse), 77 players, 29 were successes (~38 %). Those that had more strikeouts than walks, but still below the median, 25 of the 60 players (~42 %) were successes.

It seems that if we want to use empirical probability to predict the odds (percentage wise) of future players (particularly 1st basemen) having success in the Majors, we have the tools to do it. If we break the draft down a little, everything showed good correlation other than just strikeout percentage. So using the percentages in the parentheses above, we will be predicting future hitters currently in AA or AAA (and Ruf). For the draft, we will put players drafted in the 1st and 2nd round at 45% success rate, for 3rd to 5th round 31 %, and anyone taken after the 5th round is at 30 % (since 6-20 is about 30% and round 20 on is about 30%). For undrafted free agents, we just ignore this part of the formula, and divide the odds percentages by 5 instead of 6.

To give an example, let’s look at Darin Ruf. Ruf was picked in the 20th round, which gives him 30% success odds, he had above average ISO which is 41%, he falls in between the 1000 to 2000 MiLB PAs, which is 51 %, his K%-BB% gives him 38%, his walk percentage gives him 37%, and he was never a BA top 100 prospect, which gives him 29% success. Now, we add those together, which gives us 226. Then we divide by 6, which gives us 37.67, meaning according to our formula, Ruf has a 37.67 % chance of becoming an offensive success in the Majors according to our formula.

For the rest of the players, I used the following qualifications, must be 27 or younger, must be a qualified 1st basemen (according to Fangraphs’ leaderboards) at either AA or AAA in 2012, and cannot have played in the Majors (I kept Joe Mahoney in since he has played in just 2 games in the Majors). Luis Jimenez is sort of a test case since he was an obvious failure in his short stint in the Majors. I left out Matt Clark since he is signing in Japan. Even though we aren’t using strikeout percentage in the odds, I am keeping it for the “future” hitters just because it makes it easier for me to calculate K%-BB%. When you download the spreadsheet (you can do it here), you will see two tabs at the bottom, a “past” and a “future” tab. The future tab is the one you are going to want to look at to see the odds for the AA and AAA hitters (including two Rule 5 draft picks). It is sorted by the odd percentage, and again, you can download it by clicking here.

Remember, these odds do not work on lower minor leaguers and they are only for hitting and do not take into account baserunning or defense. It doesn’t measure who will make the big leagues, but tries to measure which players in AA and AAA, near the Majors, are most likely to have success. This is really basic, but maybe it will help us predict and make judgements on upper level prospects a little better. It isn’t a substitute for scouting, it is a tool to go along with it. By it’s nature, it is very pessimistic for the most part, but I think that isn’t a bug, it’s a feature. Minor League attrition rates are high. Very few minor leaguers turn out to be above league average hitters in the Majors. Also, the formula creates a heavy concentration towards the mean, around 35 %, as the worst a prospect could do (over 4000 MiLB plate appearances after being selected after the 5th round and never rated by Baseball America and striking out in every plate appearance) is have a 26.33 % chance of success and the best (less than 1000 MiLB PAs, rated as a top 100 prospect, above average ISO, more walks than strikeouts, walking over 15% of the time after being drafted in the 1st round) a prospect can do according to the formula is to have a 51.83 % chance of success hitting wise. This is somewhat of a bug, but I do like that it doesn’t guarantee success for anyone.

Tags: Baseball Statistics Fantasy Baseball MLB Prospects Off The Radar

comments powered by Disqus