Nerd-Onomics

Batters Expected BABIP

2008-05-04T18:16:00.004-04:00

Both Derek Carty and I have discussed the statistic BABIP in our columns here. Batting Average of Balls In Play measures the proportion of non-home run hits relative to all plate appearances excluding home runs, walks, sacrifices and strikeouts. We generally expect a player with an extremely low BABIP to improve as the season progresses and the adverse can be said for those with ridiculously high counts. There are exceptions, though, in situations when extremely slow or fast players hit a high percentage of their balls on the ground. For instance, if both BJ Upton and Pat Burrell hit 50 % of their balls on the ground, Upton is going to have a higher BABIP; he is faster and will leg out many more hits.

Dave Studeman, writer for The Hardbal Times, discovered a few years ago that the expected BABIP could be found by taking the line drive percentage of a batter and adding .120. If BJ Upton hits 20.5% line drives then we would expect his BABIP to be around the .325 mark. Any lower than that would indicate he has been unlucky and any higher than that would suggest luck. This is a tremendous tool when it comes to fantasy baseball because you can buy low/sell high based on what is expected to happen.

A player wants to hit as high a percentage of line drives as he can, so here are the top five line drive hitters with lower than expected BABIPs:

Edgar Renteria: 31.9% LD, .347 BABIP, .439 xBABIP
Todd Helton: 27.8% LD, .305 BABIP, .398 xBABIP
Nick Johnson: 26.9% LD, .226 BABIP, .389 xBABIP
Adrian Beltre: 26.7 % LD, .309 BABIP, .387 xBABIP
Stephen Drew: 25.3% LD, .268 BABIP, .373 xBABIP

Anybody thinking about trading or unloading Nick Johnson should hold onto him because he has been very unlucky this year. This will not keep up. Nobody is going to hit that many line drives on a consistent basis and have most of them caught. Here is the adverse side, the players with higher than expected BABIPs:

Kosuke Fukudome: 21.7% LD, .413 BABIP, .337 xBABIP
Jeremy Hermida: 19.1% LD, .394 BABIP, .311 xBABIP
Josh Willingham: 18.7% LD, .368 BABIP, .307 xBABIP
Yunel Escobar: 17.4%LD, .353 BABIP, .294 xBABIP
Torii Hunter: 17.9% LD, .348 BABIP, .299 xBABIP

Now, I am not necessarily advocating the buying/selling of these specific players but rather introducing an extremely helpful tool in determining when to trade someone. Though it does occasionally happen that a player can exceed expectations over the course of a season, if you are one to play by the odds then I would highly suggest going to Fangraphs and comparing the LD% to the BABIP. When someone comes forth with a trade this is a very interesting way to determine how to evaluate the proposal.

RBI and RBI Opportunities

2008-04-14T21:12:00.002-04:00

When it comes to offensive fantasy value one of the key statistics is runs batted in (RBIs). We intuitively equate good offensive production to the amount of RBIs a batter accrues over a season and so one of the strategies to apply when evaluating fantasy talent is figuring out which players will rack up the most RBIs. Generally we like to look at the players who show signs of persistence in racking up the "ribbies" but a more interesting approach, I would argue, is to look at which players may have the most OPPORTUNITY to knock in runs. Namely, which players have teammates with the best on-base percentages and are put in the most positions to knock those teammates in. In order to explore this I decided to look at the Baseball Prospectus RBI Opportunity statistics from this season and look at those with the most RBIs as compared to the leaders in RBI conversion percentage. The latter stat merely divides the total RBIs (excluding driving in yourself with a HR) by the total runs they could have knocked in. As of 4/13 games, here are the top ten RBI guys in baseball:

Mark Reynolds, 15
Joe Crede, 15
Carlos Pena, 13
Xavier Nady, 13
Pat Burrell, 13
Josh Hamilton, 13
Raul Ibanez, 13
Manny Ramirez, 12
Vernon Wells, 12
Jeff Franceour, 12

Now, let's look at the top ten RBI conversion percentages:

Joe Crede, 37.9 %
Angel Pagan, 31.3 %
Nate McLouth, 30.3 %
Jeff Franceour, 30.0 %
Chipper Jones, 29.0 %
Manny Ramirez, 28.6 %
Paul Konerko, 28.1 %
Jason Kendall, 28.0 %
Corey Hart, 27.8 %
BJ Upton, 26.3 %

Though the drafts are generally finished since the season is well underway if you are looking for RBI production in the waiver wire I would strongly suggest using the BP OBI % tool to find the players with the most opportunities. They may not be the best overall players available but will be put in positions to knock in runs more often than others. As you can see, some of the players listed in the raw RBI totals list are not in the OBI % list, due to a combination of generating a high percentage of RBIs through solo home runs and having many teammates on base in their PAs as compared to others.

Understanding Opening Week Statistics

2008-04-06T23:07:00.002-04:00

I have participated in numerous fantasy leagues throughout the course of my baseball fandom and, without exception, somebody in said leagues always makes somewhat drastic roster decisions based on opening week statistics. Chris Shelton hit 9 HR in 13 April Games in 2006 and numerous people rushed to the waiver wire to pick up this second coming of Babe Ruth. Unfortunately for them Shelton hit 7 more HR during the next 102 games and finished as the most disappointing/overrated fantasy player of the year depending how you look at things. While Shelton is merely one example there are always examples like this and, for whatever reason, many of us still have not learned that you cannot tell much about a player after the opening week of a season.

I am sure many people have rushed to get their hands on Nate McLouth and/or Xavier Nady but, really, do you expect to see these guys with OPS counts of above .900 or HR-RBI in the 25-95 range at season's end? If not, then why drop someone who may be in the midst of a slow start for someone overachieving now?

Though it may seem obvious the sample sizes in just 5-7 games are incredibly minute; because the statistics are not significant or stable it makes no sense to make drastic moves involving players that can only disappoint. The problem here is that we will always be one step behind. Think of it as similar to the stock market in the sense that you want to buy low and sell high; in fantasy baseball you never know which players will get off to hot starts so what happens is we buy them when they are at their peak and fail to find takers when they inevitably plummet back to Earth.

I opted not to show statistics in this post simply because I want readers to understand drastic decisions SHOULD NOT be made based on just this week. Jeff Keppinger may look great but he is filling an injury void right now; when Alex Gonzalez returns his job may be in jeopardy. AJ Pierzynski currently has the highest everything but, come on, it's AJ Pierzynski. He will be lucky to have a .265/.330/.430 by season's end.

Unless somebody has proven himself to be consistent in a certain statistic do not go crazy with only opening week numbers. Or, if you are going to base decisions this early make sure you do not drop somebody that is off to a slow start but will no-doubtedly pick it up.

There is a reason that many people will say "Well it's just the first week" when asked about their star player's slow start: it literally is just the first week. If we are quick to point that out to slow starts we should be equally wary, if not moreso, about hot starts.

Starting Pitchers: Top-Heavy or Consistency

2008-03-31T14:10:00.002-04:00

As the major league season gets underway fantasy drafts are becoming scarce along the lines of late high-school students after the bell rings; most of the students are already inside but there are a few who still want to get in. With these drafts come the differing (and always correct) strategies of every participant and expert. While most will content it is much more difficult to find a great second-baseman than first-baseman, starting pitchers will usually find themselves at the forefront of strategic arguments.

Is it better to go for top-tier pitchers a tad earlier or to wait until later rounds and end up with a bunch of good, not great, starters?

It has always been my strategy to draft guys like Johan and Oswalt early and then get #3-caliber pitchers a bit later. Others, however, have been equally successful while employing many Derek Lowe's and Andy Pettitte's, good pitchers but not of the same fantasy caliber of Johan or Oswalt. In order to investigate the different strategies I will call upon my SP Effectiveness System. For those unfamiliar the system weights twelve different statistics in order to determine which pitchers were the most effective regardless of sometimes luck induced stats such as W-L and ERA.

The SP Effectiveness database I created holds stats for every SP from 2000 until now; it also breaks the pitchers into different groups based on the rotational position they pitched similar to. Looking at these rotational breakdowns, year by year and accumulatively, helps us determine which strategy is better or if it is a wash.

With regards to specifics I am going with a 6-SP approach: the top-heavy group will consist of two #1's, one #2, and three #3's; the consistency group will consist of five #2's and one #3. The argument in favor of the top-heavy group is that the pitchers are usually proven aces that have consistently racked up a ton of relevant statistics. Of course the detriment is that this strategy prevents fantasy owners from drafting some great position players. The consistency group likes how it might be able to get equal production, later down the line, thereby allowing them to draft these better production players.

I took the averages of #1, #2, and #3 SP's and used those numbers for each year of this study. Let's start by looking at an example. For instance, in 2000, a #1 SP averaged +69, a #2 at +41, and a #3 at +31. The top-heavy group would be:

= 2(#1) + 1(#2) + 3(#3)
= 2(69) + 1(41) + 3(31)
= 272
= 272/6
= 45.3

Therefore, the top-heavy group in 2000 would average a +45.3 in SP Effectiveness Points. Here are the results for the consistency group:

= 5(#2) + 1(#3)
= 5(41) + 1(31)
= 236
= 236/6
= 39.3

The consistency group would average +39.3 SP Effectiveness Points. Let's take a look at 2000-2007 results:

2000: Top-Heavy = 45.3, Consistency = 39.3
2001: Top-Heavy = 43.3, Consistency = 40.0
2002: Top-Heavy = 43.7, Consistency = 39.3
2003: Top-Heavy = 41.7, Consistency = 39.0
2004: Top-Heavy = 42.7, Consistency = 37.3
2005: Top-Heavy = 42.2, Consistency = 39.8
2006: Top-Heavy = 41.5, Consistency = 38.2
2007: Top-Heavy = 42.8, Consistency = 40.8
Overall: Top-Heavy = 42.9, Consistency = 39.1

The results tell us that going with my strategy has only had a slightly more effective outcome in terms of starting pitching. Make no mistake: this does not mean the strategies are equal. Since both are within three points (very close) it means that it would probably be better to draft the better position players early and go for the Lowe/Pettitte as opposed to the Oswalt/Santana. If the results were staggeringly in favor of a top-heavy group then yes, get the aces, but since they are not it would probably be a better strategy to spend a 4th round pick on a great 3B or 1B instead of an ace pitcher.

Predicting Wins

2008-03-17T14:10:00.000-04:00

When I first got involved with fantasy baseball I did not have a set strategy other than knowing that pitchers would be my key to success. Since the leagues I found myself in often ended up being head-to-head, my guys could afford to have “off” weeks as long as my opponents were worse. In roto leagues this is impossible, but the strategy that helped me win three consecutive years is very sound in the logistics department. It’s also very simple – a tremendous pitching staff is much harder to create as you go than a great batting lineup and the best pitchers are the ones who are out there the longest.

While a statistic like Wins is a terrible indicator of quality it is still used in fantasy leagues; if it is going to remain prevalent then we need to find a way to get the guys who seem likely to win the most. The old creed was to just draft Yankees but now I am advocating, and strongly so, drafting the pitchers with the most innings pitched or track record of consistency with their innings pitched. Livan Hernandez is the exception to this rule as he routinely racks up a ton of innings but really has not been effective since 2004.

The key with this is to separate your sabermetrics self from the fantasy player self. Carlos Zambrano is not as effective of a pitcher as his numbers may have you think, but he is a good pitcher on a winning team that has shown a good track record of durability. While these innings eaters might not truly be the best, they will help rack up wins and solid ERA/WHIP counts. Those out there the longest have the best chance of recording decisions; and they are out there the longest, save for Livan, due to being able to post solid ERA and WHIP numbers in the individual games. Here is the correlation of IP and W over the last few years:

As you can see, those in the top five of wins usually tend to fall in the top five of innings pitched. For these reasons I would strongly suggest going after the guys that have shown a track record of innings pitched regardless of the outcome thus far. A guy like Bronson Arroyo has only gone 23-26 over the last two years but his Adjusted W-L is actually 33-16, plus he has gone over 205 innings each year. Why not draft him? You can get him in a later round, solidifying your rotation with what appears to be a bargain pick.

Though this is not some complex formula or even a new statistic, do not underestimate the value of innings pitched. When a pitcher is out there more often he is likely to record more decisions, record more strikeouts, and, again save for Livan, post lower ERA’s and WHIPS; he is out there primarily because he is not allowing runs or too many baserunners.

For those very reasons I cannot say enough about Harang and Arroyo. Roy Oswalt and Mark Buehrle are others that get supremely undervalued in a draft. While Jake Peavy may cost you a 2nd round pick, Oswalt will be just as effective and be able to be drafted four to six rounds later.

When drafting pitchers, go for the innings eaters, even if they went 12-12 or 10-14 last year. In most drafts, you're going to end up with a 12-12 or 10-14 pitcher anyway, so you might as well get the one out there the most with the best chance of recording more raw totals.

Greg Maddux: Analyzing Stats For Fantasy Baseball

2008-03-03T14:00:00.000-05:00

At my other stomping ground I recently began a series of articles profiling different aspects of Greg Maddux’s storied career. With that in mind it just made sense to profile him from a fantasy point of view. Though this article will not directly focus on Maddux, the statistics discussed will be applied to him to show his true fantasy value.

ERA is often a large part of any fantasy league. Despite this, the statistic can vary from year to year for reasons other than decreasing/increasing talent levels. The primary reason can be attributed to the DIPS theory, by Voros McCracken, that pitchers generally have no control over what happens once a ball is put into play. Walks, hit batsmen, strikeouts, and even home runs are factors controlled by the pitcher but nothing else. To read Voros’ DIPS theory, click here.

Due to his research, it was determined that, just like W-L records, ERA can be effected by luck and not just skills; in this case, how solid the defense plays, where the balls are hit, etc. Tom Tango (www.tangotiger.net) invented a statistic called FIP (Fielding Independent Pitching) that can help us understand how well a pitcher performed based strictly on the events and outcomes within his control. Tom’s formula for FIP is:

FIP = ((13*HR + 3*(BB+HBP) – 2*K) / IP) + 3.2

The 3.2 refers to a league factor used to make the FIP result look very similar to ERA. This similarity allows fans to easily relate the findings to a barometer instilled in fantasy leagues. I will use the 2004 season of Greg Maddux to show an example of FIP in action. In 2004, Greg went for 212.2 innings, giving up 35 home runs and 33 walks, while striking out 151 and hitting 9 batters. His ERA was 4.02 - the highest it had been since Maddux started his excellence in 1988. His FIP would be:

= ((13*35 + 3*(33+9) – 2*151) / 212.666) + 3.2
= ((455 + 126 – 302) / 212.666) + 3.2
= (279 / 212.666) + 3.2
= 4.51

Though Maddux posted an ERA slightly over 4.00, still respectable, his peripheral statistics were more indicative of a 4.51. Maddux had some luck on his side in 2004, which seemed to evaporate in the next three seasons. Here are Maddux’s ERA and FIP numbers from 2005, 2006, and 2007:

Over the last three seasons, his skill-based pitching was more indicative of ERA’s much lower than his actual posted ERA. A stat used by The Hardball Times is FIP-ERA, which does exactly what the title suggests – it subtracts the ERA from the FIP. If the number is negative then the pitcher was unlucky and the opposite can be said for luck. Maddux posted a +0.49 in 2004, but then posted –0.27, -0.43, and –0.57 in the following three seasons.

Carlos Zambrano and Dave Bush
In a previous article I compared the results of the Seidman SP Effectiveness System of Carlos Zambrano and Dave Bush. My system does not factor in ERA, but rather peripheral statistics, and found this comparison:

Basically, based on peripheral stats, lucky decisions, and the quality of games started, Zambrano has not been much better than Bush in the past two seasons. When we add in their 2007 ERA, FIP, and FIP-ERA, the comparison grows even stronger:

Not only have these guys been essentially equal, all things considered with peripheral stats, but their ERA’s are equally deceiving in opposite directions! Zambrano will get drafted much earlier than Bush and should not necessarily be valued as much. His luck will not continue for his entire career, just as Bush’s lack of luck will not continue forever.

2007 FIP-ERA Leaders

First, those whose ERA’s were much worse than they should have been:

1) Kip Wells, -0.97
2) Jeremy Bonderman, -0.78
3) Matt Belisle, -0.78
4) Jose Contreras, -0.74
5) Greg Maddux, -0.57

Contreras makes sense being here as his 2007 looks like an outlier compared to the rest of his seasons. I would fully expect him to post solid 2008 numbers mainly because his 2007 was just unlucky. Maddux is the only player on this list whose ERA was respectable coming in. The other four still had high FIP’s.

Next, those who got aided by luck:

1) Fausto Carmona, +0.99
2) Matt Chico, +0.90
3) Livan Hernandez, +0.85
4) Jeremy Guthrie, +0.83
5) Oliver Perez, +0.80

The odd thing about Carmona is that his ERA gets worse when the context of his actual skills are called into play, however, with my Adjusted W-L Record system, his W-L record gets better. Using regular W-L and ERA, Carmona was 19-8 with a 3.06, but using the adjusted numbers of Adj W-L and FIP, Carmona was 23-4 with a 4.05 FIP.

When trying to find pitchers to round out your rotation, use FIP over ERA, as well as my Adjusted W-L Record system over the regular W-L to find those whose skills greatly outweigh their luck or lack of luck.

BABIP For Batters

2008-02-18T13:56:00.000-05:00

Derek Carty, a colleague of mine, wrote an article not too long ago revolving around BABIP and the DIPS theory from the pitcher’s point of view. My goal in this article is to analyze the BABIP statistic from the point of view of the batter. To read Derek’s article, click here.

BABIP stands for Batting Average of Balls In Play. The statistic can be calculated with the following formula:

BABIP = (H – HR) / (AB – K – HR + SF)
The statistic measures the percentage of balls in play that fall in for hits. Since home runs are considered to be a pitcher-dependent statistic, BABIP removes them from qualifying. Strikeouts are also considered to be pitcher-dependent and so this statistic primarily measures the amount of times a batter gets a hit out of the amount of times he hit the ball. For an example we will look at the 2007 season of Philadelphia Phillies left-fielder Pat Burrell. Burrell had a BA (Batting Average) of .256 while going 121-472 in 155 games. Burrell hit 30 home runs, struck out 120 times, and had 8 sacrifice flies. Plugging him into the BABIP formula:

Burrell BABIP
= (121 – 30) / (472 – 120 – 30 + 8)
= (91) / (330)
= .276

In 2007, Burrell had a batting average of .256 but a batting average of balls in play of .276. This tells us that Burrell’s average took a big hit (almost 20 points) due to his high strikeout count. Essentially, since the stat removes home runs from each side, BABIP tells us what a player’s batting average would be if he never struck out. The reason this statistic can be useful deals with the context in which it is used. We should not, by any means, replace BA with BABIP. Players do strikeout and so using BABIP as a barometer would not be a fruitful way of analyzing talent. What BABIP can tell us is which players would be good or bad fantasy picks.

The key, above all else, is consistency.

Make no mistake, if a player posts consistent BABIP regardless of his BA, then we are dealing with luck and factors seemingly out of his control. These types of players are the ones we should be looking into drafting. Of course, this really does not apply for the early rounds when we all know which batters are going to be drafted, but rather for the later rounds, when we are deciding between the Felipe Lopez’s and Ryan Church’s of the world. Those who show extreme fluctuations in BABIP are the ones who get drafted higher the next year and then usually are used as trade bait or deemed busts when they do not match their previous year’s success. The problem here is not that the player was a “bust” but rather that the owner that drafted him did not do his homework and was instead WOW’ed by flashy numbers.

Again, let’s look at Pat Burrell, this time analyzing his seasonal BABIP since his MLB debut in 2000:

This table tells us that Pat Burrell, other than one awful season, has consistently had very respectable BABIP numbers. Though he appears to be on the decline since 2005, you can see that his BABIP has been pretty consistent despite the fluctuations in his batting average. These are the types of guys that are good bets for fantasy picks due to the consistency in their percentages of balls in play that drop in for hits.

I’m not telling you to draft Pat Burrell, or draft him high, or anything along those lines. What I am saying, though, is that when it gets down to the final few rounds of a draft and you need to fill out your team, incorporate BABIP into the equation. If you’re having trouble deciding between a Juan Pierre or Dave Roberts, guys that will rack up steals, go with whoever has had a more consistent BABIP throughout the recent years.

BABIP will also progress or regress, usually towards a .300 range. This can be seen in Burrell’s table. In 2000 and 2005 his BABIP were abnormally high, however his very low 2003 makes up for it.

The entire point of this article, in suggesting you apply BABIP to your fantasy evaluation arsenal, is to be wary of breakout seasons and incredibly “down” seasons. Most would shy away from a player like Burrell following the 2003 and 2004 seasons due to his terrible 2003, however he was essentially the same type of player in 2004 as he was in 2001 and 2002. Before getting very high on players who come out of nowhere to lead the league in BA or very low on those who have significant dropoffs, do yourself a favor and investigate their BABIP – it will tell you whether or not this was a fluke or something to look out for.

SP Effectiveness System: The Fantasy Application

2008-02-11T13:38:00.000-05:00

Last week I finished my database and turned all of my research, explanations, and statistics on evaluating Starting Pitchers into an 83-page register that Brad Stewart was gracious enough to post here at MLB Front Office. The register holds career and seasonal statistics dating all the way back to 2000 and can greatly aid your fantasy team. The lone problem I foresee with how it may be able to help you is that, unless you understand how and why the system works, the register can be pretty intimidating.

My whole goal in sabermetrics has always been to provide my readers with accurate and poignant research and explanations while breaking everything down so that even the least stat-savvy person can understand what I am discussing. Due to this, I felt compelled to explain the register a bit and how you can apply it to your fantasy baseball strategy in order to build an extremely solid starting rotation.

The register can be found here at MLB Front Office as well as my homepage, www.ericjseidman.com.

I have to thank Jeffrey Panehal, a loyal reader of mine, for bringing this topic up. Jeff sparked the idea of parsing the data to find those who are steadily improving to aid him with his fantasy team, the Fort Wayne Newtons.

Applying the SP Effectiveness System
Since my system takes into account a wide array of statistics (most of which are counted in fantasy baseball leagues) those who earn the highest number of SP Effectiveness Points are generally the best starting pitchers to have in a given year. Now, the system is not truly tailored to work during the season yet, but in this article I am going to provide you with results from 2005-2007 based on some observational tendencies I made.

I am not, by any means, going to tell you who to draft or when to draft him but I am going to break down the starting pitchers into different categories based on the findings in my database from the last three years.

Different SP Categories
Not every starting pitcher is going to find himself in one of these categories. Those not included, simply put, showed no upward or downward trend. For instance, Jason Marquis posted a +34 in my system in 2005, a –2 in 2006, and a +19 in 2007. There really is no trend or tendency there. He was as effective as a high-end #3 starter in 2005, ineffective in 2006, and a #4 starter in 2007. I’m not saying that he is a bad pitcher or one that you should completely avoid but he will not be included in this article (other than right now) because he has shown no sign of consistency (positive or negative) since 2005.

That being said, I am breaking the SP’s into six different categories: Steady #1’s, Steady #2’s, Total Pts Increasers, Tot/Gm Increasers, Total Pts Decreasers, and Tot/Gm Decreasers. The last four categories refer to those who have seen their numbers in the titled statistic increase or decrease during all three seasons (or both seasons if they only pitched in 2006 and 2007).

The Steady #1’s category refers to only those who have posted an NL +50 or higher or an AL +47 or higher in their seasons during that span. The Steady #2’s category refers to those who have posted an NL +36 to +49, or better, or an AL +34 to +46, or better.

I have also included a few pitchers in this category that posted two #1 or #2-caliber seasons from 2005-2007. As long as the non #1 or #2 season came in 2006, and was sandwiched by two better seasons, they qualify for this category. The major example is Mark Buehrle.

Buehrle has posted a +45 or higher in every season of his career except 2006. In 2005 he posted a +56, in 2006 he went all the way down to a +18, and in 2007 he shot back up to a +59. The fact that he has been so consistently good in his career and jumped all the way up to such a high score in 2007 tells me 2006 was most likely a fluke, thereby qualifying him for this category.

Another thing to keep in mind is that guys with very high scores are not included in the “decrease” statistics. Johan and Halladay have decreased from 2005 to 2007 but their scores have been so high that this statistically proves what people have said – that a down year from one of these guys is still better than an up year for most others.

The positive categories (everything that does not have the word “decrease” in it) consist of players you may want to look out for. I highly suggest you download my register and refer to the Effectiveness Points each year for those who have increased. Some will increase but still be at a low total. IE – Daniel Cabrera has actually increased his Eff. Pts and Eff. Pts/Gm in 2005, 2006, and 2007, but his total SP Effectiveness Score is still relatively low despite the improvement. The SP Effectiveness Scores will be shown progressing from 2005 to 2007.

Total Pts Increasers
Keep in mind that these players increased from 2005 to 2006 AND 2006 to 2007, or if they only pitched two years, increased from year one to year two. I did not include players who decreased from year one to year two but increased from year two to year three.

Tot/GM Increasers
The same rule from the previous category applies – must increase every year.

Total Pts Decreasers
The same rule from the previous category applies – must decrease every year.

Tot/GM Decreasers
The same rule from the previous category applies – must decrease every year.

If you want to use these results to find the players that are showing the most positive upward trends, find the highest total on the Total Pts Increasers who is the highest on the Tot/GM Increasers. In this case, Cole Hamels would fit that bill. Hamels increased his total score by +48 points and his Tot/GM by 1.58. Gil Meche, CC Sabathia, and James Shields would come next. Again, this is not a suggestion as to who the best pitchers in the league are, as those would be in the Steady #1’s or Steady #2’s but these Increasers and Decreasers categories show us which pitchers are steadily improving or declining.

To determine which players to stay away from refer to the register to see their year by year scores. The reason I suggest that is because some players show decline in their effectiveness due to having such a good 2005. Dontrelle had such an effective 2005 that it was very hard to improve upon and so his decline, while vast, has to be partly attributed to how good his 2005 turned out.

There you have it. This is the best way to apply my register to fantasy. You could also just look at improvements from 2006 to 2007 to help filter results. The register is vast and contains a ton of statistics and data.

Another way to utilize it would be to examine which pitchers are consistently unlucky (better than barometers) or lucky (worse than their barometers) as a way of predicting improvement.

Adjusted W-L Record

2008-02-04T13:34:00.000-05:00

In the world of fantasy baseball a win is a win and a loss is a loss. Regardless of how bad a win may be or how good a loss may be, the above statement will seemingly always hold true. Fantasy leagues do not care about Cheap Wins and Tough Losses, but it is my strong conviction that you should care.

With that in mind I would like to introduce you to another statistic of mine – The Adjusted W-L Record. Since W-L records tend to indicate luck instead of skill, Adjusted W-L gives us a tangible record that truly measures the performance level of a pitcher. No-Decisions are not counted so the Adj. W-L is designed to show skill relative to decisions received. It answers the question most of us want a regular W-L to answer – What would happen if a pitcher got a win for every well-pitched decision and a loss for every poorly-pitched decision?

W-L Breakdown
A W-L record consists of Cheap Wins, Tough Losses, Legit Wins, and Legit Losses, all of which revolve around the AQS. If you record an AQS and win it is a Legit Win. Recording an AQS and losing is a Tough Loss. Failing to record an AQS and winning is a Cheap Win and failing to record an AQS and losing is a Legit Loss.

Adjusted W-L isolates the cheapies and toughies from the legitimate decisions and then re-inserts them backwards. Cheap Wins become losses and Tough Losses become wins. If W-L records are going to be a barometer of quality, we might as well use a version of it that actually tells us about quality.

In 2007, Josh Beckett went 20-7 and mostly everybody loved him. Despite such a solid record he finished #8 in the AL in 2007 SP Effectiveness System. How did this happen? Well, it wasn’t that Beckett was bad, because his 20-7 would barely be adjusted to a 19-8, but the guys listed higher than he had better Adjusted W-L.
We cannot help ourselves, when seeing one, to equate a W-L record to quality. It is assumed that Beckett, at 20-7, was better than Dan Haren and his 15-9 record. Beckett’s 20-7 gets adjusted to a 19-8, but take a closer look at Haren. Most will say he performed better than a 15-9 record would indicate but have trouble pinpointing how much better.

Haren had 0 Cheap Wins, 15 Legit Wins, 3 Legit Losses, and 6 Tough Losses. Using the formula his Adjusted W-L in 2007 would be 21-3. Fausto Carmona’s 19-8 becomes a 23-4. Mark Buehrle’s 10-9 turns into a 15-4 and John Lackey’s 19-9 was actually closer to a 22-6. If all of these guys posted the records they should have posted, based on how well they performed, do you think there would be any problem or odd reaction to seeing Beckett below them on my system? I tend to think not.

Fantasy Application

Even though fantasy leagues do not differentiate a Cheap Win from a Legit Win, and so forth, I have found that those with the best Adjusted W-L records usually post the best peripheral statistics (WHIP, K:BB, IP, IP/GM).

Here is a PDF file showing the Adjusted W-L data of the NL pitchers in 2007.

Looking at Adjusted W-L records not only tells you who the best guys were at reaching certain statistical plateaus that are counted in fantasy leagues, but it also paves the way into the field of luck.

Net Luck Rating
An interesting way to use this involves looking at how lucky or unlucky a pitcher has been over a period of time. We can find a pitcher’s Net Luck Rating by subtracting his Tough Losses from his Cheap Wins, and adding that to ½ the difference of AQND’s from Non-AQND’s. As a quick refresher, an AQND is a AQS with a no-decision and I think you can figure the other one out.

Carlos Zambrano, in 2007, went 18-13. He had 0 Cheap Wins and 2 Tough Losses. He only had 2 No-Decisions and both were Non-AQND. His Net Luck Rating would be:

= (CH. W) – (T.L) + 0.5(Non-AQND – AQND)
= (0) – (2) + 0.5(2 – 0)
= -2 + 1
= -1 NLR

Zambrano’s 2007 Net Luck Rating was –1, meaning he was just slightly unlucky. Negative numbers indicate the pitcher was unlucky and positive numbers reflect luck. In 2001, Roger Clemens won the Cy Young Award with a 20-3 record. He had 6 Cheap Wins, 0 Tough Losses, 6 Non-AQND, and 4 AQND. His NLR would have been:

= (6 – 0) + 0.5(6 – 4)
= 6 + 1
= +7 NLR

In other words, Clemens was very lucky.

In case you are wondering why the No-Decisions are multiplied by 0.5 the reason deals with the fact that No-Decisions, while preventative of the pitcher receiving the correct decision, do not adversely effect the W-L record, whereas Cheap Wins and Tough Losses do.

Conclusion
My database of starting pitchers is complete and it dates back to 2000. I am currently turning it into an official register/guide. The PDF document will be free and available at MLB Front Office as well as my personal website, www.ericjseidman.com.

The register includes year-by-year statistics for all pitchers in the database and I plan to make this an annual “publication.” Stay tuned as it will be complete before the end of February. If we want to use W-L records as a barometer for fantasy quality, use my Adjusted W-L since it shows the skill, not luck, that we seek in a W-L record.

Introduction

2008-01-28T13:29:00.000-05:00

Welcome to “Nerd-onomics,” a weekly Sabermetrics column to be run by me, Eric J. Seidman. Instead of jumping right into my first column I want to properly introduce myself. I want to explain who I am, what I have done, and what you will get from me. I have also included some brief explanations on certain statistics, formulas, or systems I created and like to use.

Who Am I?

My name is Eric J. Seidman and I am a screenwriter, sabermetrician, and magician from Philadelphia. At only twenty-two years old I have already won best screenplay at a film festival and had three scripts optioned by well-known producers. I was able to get Brett Ratner (director of Rush Hour, The Family Man, Red Dragon) to executive produce a short film of mine and I somehow got hired to write the story of Not Another Indian Movie.

I have worked my tail off to earn every inch of success, which serves as a testament to my meticulous work ethic and exponentially upgrading determination. The reason I decided to mention all of my movie success hinges on the fact that movies are not even my truest passion. That honor goes to baseball. As my bio states, by virtue of growing up in a sports-heavy household, I was the only second grader that immediately knew a fraction of 7/19 equates to a .368 batting average. If I was able to accomplish so much in a field that is not even my true passion, imagine what I could do if I applied my work ethic, determination, and knowledge to an area I love more than anything other than my girlfriend.

In addition to writing here at MLB Front Office I also write for MVN - Most Valuable Network and I will likely be contributing to The Hardball Times. At eHow.com I am the Magic & Performance Expert where I specialize in how-to tutorials revolving around magic tricks and performance. I am in the process of writing my first book on Sabermetrics, titled Bridging the Statistical Gap. The book will cover all of my research, theories, and statistics, and will be written in a style that even the least stat-savvy fan can understand. Now that we are formally introduced let’s get into what you, as fantasy players and readers, can expect from me.

What To Expect

My goal in writing here at MLB Front Office is to be your personal advisor. I am going to statistically show you ways to enhance your fantasy team as well as ways to find similar production levels from players you might not expect. I want you to have the best team, without over-spending, and be able to find quality stat-machines from a garbage bin or scrap heap.

My e-mail address is seidburns850@aol.com and I encourage you to e-mail me whenever you want. I answer every e-mail I receive, in very timely fashions, and am here to serve your fantasy baseball needs in any way I can. My phone number is 267-255-6777 and, while I would prefer e-mail, feel free to call, text, or leave a voice-mail.

Up until the regular season starts I will be writing every Monday about my statistics and evaluation tools, being sure to provide results and concrete evidence. I absolutely hate the statistical barometers that are being used today, since they do not tell us even half of the whole picture, and so I am going to show you better ways to determine quality and effectiveness. After the season gets underway my focus will shift to what is actually happening in the current season.

I am not going to tell you how to draft, who to draft, or when to draft him, as my esteemed colleagues here do an insanely tremendous job of that. What I will do, and you have my pledge here, is find quality production out of players that most will spit on or not even consider. For instance, according to my Starting Pitcher Effectiveness System, Dave Bush was almost just as effective as Carlos Zambrano in 2006 and 2007. In 2006, Zambrano had a +45 and Bush had a +43, and in 2007, Zambrano had a +35 while Bush had a +31. Zambrano gets drafted high, based on his reputation, and most do not even think about Bush even though in the last two years he provided just about equal production.

Seidman Starting Pitcher Effectiveness System

I have turned all of my research in this weighted points system into a handy-dandy PDF file for you to look at. To sum it up, though, the system takes into account a plethora of variables and truly levels the field of play between guys on good/bad teams, guys with/without run support, and guys injured/called up as opposed to just plain bad. I strongly suggest you read the PDF file because I am going to reference it as soon as next week when dissecting certain bargain bin pitchers who, over the last few years, have been much more effective than their barometers (W-L, ERA) would indicate.

The PDF file can be downloaded here.

AQS – Adjusted Quality Starts

One of the biggest parts of the SP Effectiveness System is a statistic I created called Adjusted Quality Starts. Normally, Quality Starts are defined as games in which a pitcher goes for at least six innings and gives up no more than three earned runs. Well, what if a pitcher goes eight or nine innings and gives up four runs? Isn’t that an equal, or better, ratio of earned runs to innings pitched? It is, but despite that, it would not be counted as a Quality Start.

The AQS extends the rule of 6+IP and 3 or less ER to also include games of 7.2+IP and 4 or less ER. Due to the increasing role of the bullpen and the severe drought of pitchers reaching the eighth inning, if you can go 7.2 IP or more, and give up no more than four runs you will earn an AQS.

This statistic will be extremely important for fantasy owners because it evaluates every single game of a season, not just the whole season. If we see that Pitcher A, in 3 games, went for 20 IP, 15 H, 8 ER, 8 BB, 20 K, and average it out, he would have a three-game average of 6.7 IP, 5 H, 2.7 ER, 2.7 BB, and 6.7 K. Those are great numbers, but they are only averages. Those numbers would indicate that the pitcher had three good starts. Sometimes that will be the case but not always. What if he went for nine innings and no earned runs in game one, but went only five or so innings surrendering four earned runs in games two and three? He would have the same averages even though he only had one good start and had two bad starts. In fantasy leagues, this is a big difference since a pitcher will not post his entire season’s statistics in one week of a head to head league. The AQS measures how often a pitcher was good, not necessarily how good he was in those good starts. Wouldn’t you rather have a guy who has a higher percentage of good starts as opposed to one with 50 % great starts and 50 % bad/ineffective starts?

Next Week

This week I merely wanted to introduce myself so that, in the future, you didn’t just randomly see some new guy posting baseball information. Now that you know who I am, what I like to do, and what I plan to do, let’s briefly discuss what you will get next Monday.

Next week we will discuss Starting Pitchers. I will be using my SP Effectiveness System as a major referencing point when discussing the starters so be sure to either know the system, or at least have it printed to understand what I am talking about. We are going to take a close look at the following areas –

1) SP’s with big reputations and stats that back them up
2) SP’s with big reputations and the stats do not back them up
3) SP’s with no real reputation but stats similar to those in #1
4) SP’s with no real reputation and stats that would not disagree
5) SP’s with bad reputations who are better than you think

Have a great week and, as always, feel free to e-mail me with questions, concerns, or anything along those lines. I’m glad to be onboard and hope I can be of some help to you.