By Eriq

A few weeks ago, I pointed out that pitching has gotten stronger this year.

One of the main explanations is that batters are hitting fewer home runs. Last year, across the major leagues, a HR was hit every 32.9 at bats. This year, a HR is hit every 36.4 at bats. I’m not sure if that sounds like a lot to you, but it’s about a 10% decrease in total number of HRs.

Now this has some very subtle but pretty interesting implications for fantasy baseball.

If you put particular attention on speed during your drafts this year, thinking it was a scarce commodity, suddenly at least relative to power, it’s a little less scarce. The ratio of HRs-to-SBs this time last year was 1.55. It’s now down to 1.5. This means that the value of a single HR has gone up slightly and the value of a single stolen base has gone down slightly.

The number of players at each HR total in 2009

On the other hand, we’re  not necessarily sure this makes sluggers themselves more valuable. Yes, the HRs they hit are more valuable….

But for two reasons, having a player who is on the HR leader board might mean a little less.

For one thing, there’s the distribution of HRs.

Think about it. If less HRs are being hit, the gaps between the haves and the have-nots are going to be smaller. Having a “slugger” only gets you so far.

The number of players at each HR total in 2010

At this point last year, Adrian Gonzalez led the league with 22 HRs and there were 18 guys with at least 13 HR. These players had a pretty good edge on the average player. Among players with at least 100 AB, the average hit a total of 5.5 HRs. Thus, Gonzo had a 17.5 HR edge over players who got regular playing time.

This year, Jose Bautista leads the league with 18 and there are only 8 guys with at least 13 HRs.Among players with at least 100 AB, the average hit a total of 5.2 HRs. The margin is shrinking.  Bautista has about a 13 HR edge over players who get regular playing time.

In other words Gonzo and the top sluggers of 2009 were outpacing their peers to a more significant extent than Bautista and the top sluggers of 2010 are doing right now.

The second reason why having a slugger may not mean as much is the changing nature of run production. Less HRs means that HRs aren’t accounting for as many RBIs and Runs as in past years.

Right now, if you look at the 35 top leaders in the HR category (about 10% of all those with at least 1 HR), they have together a total of 1,131 runs and 1,232 RBIs. That’s 15.4% of all total runs in baseball and 17.6% of all total RBIs in baseball.

Compare it to last year, when the 35 top leaders in the HR category at this point of the season had 16.4% of all total runs and 18.9% of all total RBIs.

In other words, being a premier slugger translates to a smaller percentage of runs and RBIs compared to their usual share.

Put this all together and what does it mean?

Here’s my theory: The value of a stolen base is down. The value of a slugger is down too. Standings should be a little tighter in fantasy leagues across America. And get this, I think that having hitters who have excellent batting averages are more important than ever.

I’ve been looking at player raters and noticing that there are more great-average players at the top, that the top players have an average that’s a higher number than in previous seasons, even though batting average in baseball as a whole has slipped from .262 to .258. You want to know why Miguel Cabrera, Robinson Cano, and Justin Morneau are the three most highly rated players in baseball right now? Sure, the good pop helps. But having an average at .355 or above is really damn special right now.

By Eriq

Individual BABIP vs. Team BABIP

By now, it’s common wisdom that pitchers can get lucky or unlucky based on opposing batters’ hit rate. We all look at stats like BABIP and FIP to determine whether a pitcher really deserves his ERA or whether a few extra balls hit into play are going for outs or hits than normal.

But that formula ignores one key aspect: Fielding does play a part in the game of baseball.

So, for example, let’s take the case of Barry Zito. His BABIP currently sits at .242. This screams LUCKY. However, the Giants lead the majors in team UZR, or ultimate zone rating. They have a pretty good defense. As a team, the Giants have a pitching BABIP of .268. So should we measure Zito by the 58 point differential between his own BABIP and the standard .300 norm or should we measure Zito by the less scary 26 point differential between his BABIP and the team’s BABIP? Perhaps Zito is a pitcher who appears to be getting lots of luck when he’s really getting a little luck combined with a lot of defensive support.

It’s nice that sabermetrics guys have extrapolated what a pitcher’s ERA should really be based at core skills like strikeout-to-walk rate and HR rate and by taking defense out of the equation, but perhaps someone needs to put the dots back together again. Namely: How does defense influence a pitcher’s ERA?

In the meantime, let’s look at players who may be appearing to get lucky or unlucky when a comparison between their personal BABIP and their team’s BABIP shows otherwise.

First, here are some pitchers with a BABIP over .300 (usually signifying poor luck) who play on teams with an even higher BABIP. These are the guys who may seem to be getting unlucky when perhaps they play for teams with miserable defenses and therefore, may actually be getting fortunate on balls hit into play!

If you look at an extended list, you’ll see a lot of Brewers, Angels, Dodgers, and Astros in the above class. No surprise those teams rank at the bottom of team UZR too. People may assume that pitchers on these teams have inflated BABIPs and will get better as the season moves along, but perhaps it just ain’t meant to be on teams with miserable fielding.

Now for pitchers with a BABIP below .300 who play on teams with fairly miserly BABIP. These are the guys who may seem to be getting fortunate when perhaps they just play for teams with tremendous defenses.

Again, if we look at a fuller list of players in this category, there will be a lot of pitchers on the Padres, Giants, Athletics, Yankees, and Rays. Yes, those teams are all among the elite class in team UZR as well.

The Tampa Bay Rays are a prime example of a team with a pitching squad who all seem to be getting lucky. If you sort the tables on Fangraphs by ERA-FIP, you’ll see pitchers like Wade Davis, Jeff Niemann, Matt Garza, and David Price all listed as having an ERA that’s “lucky” by at least a run. Perhaps it’s possible that all four of them are getting lucky, that the team’s pitching as a whole is getting lucky too with a BABIP of .268. But considering the evidence, there’s certainly a great case to be made this is no accident.

By Brian Mills

In my last post I discussed how players differ in producing HRs from week-to-week. This is important because if you play in a H2H league, you want to know whether trading for a slugger will really increase your chances of winning the HR category in a given scoring period.  Now, that I’ve looked at individual players, it’s time to analyze how HR distribution across a season effects an entire fantasy roster, given the many variables.

If you recall from my previous post, we found that we’d expect David Dejesus to at least equal Albert Pujols in Home Runs about 30% of one-week scoring periods.  But this extreme example does not take into account where your own team ranks in power relative to the rest of the league, or how the variability of the rest of the players on your team can affect the likelihood of a win.  That’s what I’ll try to do here.

Continue Reading »

By Brian Mills

If you replaced David DeJesus with Albert Pujols in your lineup, how much more often would you win the HR category in a H2H league?

It’s about a month into the season and many owners are probably worrying about whether or not their team has enough power to contend. It’s a little early to truly be worrying about slow starts, but it’s always good to plan for the future just in case.

In Roto leagues, finding a power boost is pretty straightforward: If you trade for a guy who between now and the end of the season will likely hit 15 HR more than the one you have starting now, your HR total will probably simply go up by 15 HR. It’s easy to then check the standings to see how many points those extra 15 HRs are worth and measure the benefit against whatever you’re giving up to make the trade.

However, in Head-to-Head leagues, the calculation isn’t as so pat. Sure, improving your HR total by 15 should help you to win the HR category more often in your league. But how often?

Over the course of the rest of the season, an extra 15 HR is about 1 HR per weekly session (depending on the length of your regular season). So it first depends on where you stand in HR right now. Have you been losing the Home Run category every week by 3 or 4? Or are you just on the brink of winning it each week?

Leaving aside the correlation of HR-hitting with Runs and RBI, we’ll try to figure out how much of a power boost you’ll need to win a few more H2H matches and set yourself up for the playoffs.

Continue Reading »

By Eriq

A reader wrote in to ask whether or not the numbers support sitting a starter in the first week of the season.

I understand the reasoning behind the strategy: In the first week of the season, starters may be dangerous as they haven’t built up the stamina to go long into games, may not have fully gotten control of their arsenal of pitches, and might be expected to be less likely to put up wins and more likely to damage a fantasy team’s pitching ratios.

But let’s look at the numbers.

Unfortunately, we couldn’t find any “first week” splits anywhere, but we were able to gather together data for pitchers who made starts between April 5 and April 12 going back five years. We can call this set of data, “Starters in Early April,” which may be better anyway since what we’re talking about is a pitcher’s seasonal maturity at this early part of the year. We compared this data to how starters performed overall in the past five years.

The results of the study were somewhat surprising.

Let’s start with endurance and the potential for wins. We’re shocked to learn there’s hardly any difference at all. Starters in early April average 5.8 innings per game started. Starters average the same 5.8 innings per game started throughout the season. In early April, starters win the ballgame 34 percent of the time. Throughout the season, that only ticks up to 35 percent. In early April, starters are a little bit less likely to be on the hook for a loss and a little more likely to be given a no decision, but unless your league counts those stats, it’s not very important.

Let’s go to ERA and WHIP.

Here we find big differences but in the complete opposite direction we expected. Starters in early April average a 3.92 ERA and a 1.34 WHIP. Over a full season, starters average a much worse 4.5 ERA and a 1.39 WHIP. Is it because teams mostly have their best starters healthy at the onset of the season? Perhaps that’s one factor, but I think we can explain the difference better by jumping into the peripheral numbers.

First, we find no command issues. Both time frames yield an average of 3.1 walks per 9 IP.

Interestingly, despite the better surface ratios, pitchers at the beginning of the season strike out less batters. In early April, pitchers strike out 5.5 batters per 9 IP. Throughout the season, the number jumps to 6.3. If starters are whiffing fewer batters in early April, how are they managing to gain a better ERA?

Alas, we can put the mystery to bed by taking a look at the HR numbers. In early April, the HR/9 rate of a starter is only 0.94. Throughout the season, it’s 1.1. Clearly, the biggest advantage that a starter has at the start of the season is the cold weather. If you’ve ever swung a bat in frigid temperatures, you should know it stings upon contact. Moreover, because warm air is less dense than cool air — ask your local meteorologist — balls travel further in those warm summer months.

So it may make sense to watch your starters take the mound a few times before throwing them out for service to your fantasy squad, but the numbers seem to suggest you might be missing out on the best part of their season.

An analysis with a surprising conclusion

By Eriq

Miguel Cabrera: A near-.400 OBP may not be as valuable as you think

Everyone who participates in a league that counts OBP instead of AVG knows that the format heavily rewards players like Adam Dunn, Jack Cust, and Nick Swisher who do three things well — hit HRs, walk, and strikeout. Failing to hit the ball into play and having a low batting average isn’t penalized in OBP formats if a player can take a heavy number of bases on balls.

Who are the players whose fantasy value suffer when going from AVG to OBP formats? The answer can be surprising.

Of course, players who get dinged in value are ones who hardly take a walk. We’re talking about Bengie Molina, Robinson Cano, Miguel Tejada, and Jose Lopez.

But they aren’t the only ones whose value suffer. Would you believe that a player like Joe Mauer who consistently has an OBP over .400 is less valuable in an OBP format? The same is true for Ichiro Suzuki, Pablo Sandoval, Miguel Cabrera, and Billy Butler — players who are expected to post very nice OBPs this coming season.

Here’s the problem: These players all have better advantages against their peers in batting average than they do in on-base percentage. A player like Miguel Cabrera is the 5th ranked player according to CHONE projections in a league that counts AVG, yet only the 13th ranked player in a league that counts OBP. Most competitors probably won’t realize that when looking at a player who has gotten on base nearly 40% of the time in four of the last five seasons. But it’s true. From a statistical standpoint, his batting average is more extraordinary.

Furthermore, a player doesn’t necessarily have to have a great OBP to be rewarded in OBP formats. Take Dan Uggla as an example. At best, he’ll post an OBP that’s fairly mediocre. Being mediocre isn’t totally bad. Especially compared to the alternative of being one of the league’s worst in the category of batting average.

Let’s stack the winners and losers. Below are our rankings of who benefits and who loses most when going from a league that counts AVG to a league that counts OBP. Since most player rankings are geared towards AVG-format leagues, it’s worth considering.

Continue Reading »

Can you figure out where a player is likely to be drafted without looking at ADP?

By Eriq

We recently proposed a new metric called xDP, or expected draft position, based on factors like recent production, growth, upside, risk, and scarcity.

Paul Singman at The Hardball Times has done yeoman’s work on the topic by measuring how recent production correlates with draft position. Not surprisingly, he finds that home runs and stolen bases are most important when it comes to a player’s draft position. Singman also proposes a formula to estimate a player’s draft position based on prior year’s stats — a nice leap in our understanding of the dynamics of drafts. We still would like to incorporate growth, upside, risk, and scarcity into the equation, but this is a very interesting start.

The article also notes that runs are the category given short shrift in a competitor’s drafting inclinations, which is interesting, because as we’ve noted in the past, runs have the highest single category correlation to a fantasy team’s hitting success. We agree this represents a great potential inefficiency in the market.

Who are the best keepers for 2010 Fantasy Baseball?

By Eriq

Fantasy baseball enthusiasts are gravitating in ever greater numbers to a format of play known as “keeper leagues.”

These leagues attempt to come closer to giving competitors the experience of running a real-life baseball team by allowing fantasy owners to retain a certain number of players owned in the previous season. By doing so, leagues foster active participation throughout the year and let teams build long-term strategies of success.

Picking keepers is not an easy task, though.

A couple weeks ago, on this website, we began to formulate a more quantitative approach to finding the best keepers in fantasy baseball by looking at the players who had made the most significant jumps in draft position from last year to this year. The premise is that if a team owner has a player locked in at his ‘09 value, a big draft position jump signifies a value premium over the original investment.

However, we quickly realized there are two big flaws to this approach. First, not all draft position jumps are equal. For example, Player A who moves from the 9th round to the 2nd round may have increased his worth more significantly than Player B who moves from the 19th round to the 5th round. The latter player may have made a larger jump, but Player B may have made a more valuable jump. Second, the model is beholden to the whims of the crowd, and while there may be wisdom in how those in mock drafts are going about their business, it may also be helpful to judge the value of a keeper based on some non-subjective measure of expected production.

Thus, we’ve come up with a new model. Without further ado, here are the top 10 keepers for 2010 fantasy baseball:

1. Mark Reynolds
2. Aaron Hill
3. Michael Bourn
4. Matt Kemp
5. Justin Upton
6. Adam Lind
7. Joe Mauer
8. Alex Rodriguez
9. Kendry Morales
10. Zach Greinke

You can also see a ranking of the top 100 keepers in fantasy baseball for the 2010 season. The list includes a “Keeper Power Score” for every player. We also have a link to the rankings on the left-hand sidebar.

How did we come up with the ranking and the power score? Read on…

Continue Reading »

This is a guest post by reader Brian M., who showcases his method for evaluating fantasy team performance in his H2H league irrespective of standings. The article explores use of a composite-Z score to compute a power ranking in a fantasy league. We think this is an interesting idea whose applicability into other leagues might be adjusted slightly and adapted for self-evaluation.

The end of the fantasy season can leave everybody but the champion feeling unsatisfied. 

Sometimes, as fantasy players, we suspect we performed much better than the standings seem to tell us.  In every fantasy season, there are factors we can’t control, stuff that influences the outcome of a season beyond putting together a great core of players. For example, being lucky enough to have a soft schedule with good matchups against other competitors can have a huge influence. In H2H leagues, we often strategically set lineups in chase of short-term goals—or optimal performance in short scoring periods— and the final standings reflect nothing more than a collection of small sample sets.

Maybe you suspect your team was better than its fate. How can you persuade your peers that you were really handed the short end of the stick?

Each year I try to find alternative ways to evaluate how my team fared against everyone else, independent of H2H standings. The actual standings are, of course, important above all else, but I want to know if my roster-building strategy is on the right track.  It’s easier to evaluate your teams inefficiencies in standard rotisserie leagues and points-based H2H leagues, but unfortunately, in a H2H league that implements rotisserie-style scoring—or 1 “win” for each category victory in each scoring session — trying to figure out where you stand as a fantasy general manager can be cloudy.

The table below  is an example of a 20-team H2H League with 8×8 categories and 10 regular-season scoring periods (two weeks each) . Wins—or “true standings”—are in the second column.  The league is made up of four 5-team divisions with a schedule that requires each team to play all other teams in its own division twice, along with 2 non-divisional games.  This can make for some serious skew in the standings based on divisional strength and luck. (Click through for a larger image)

The table is color-coded by division and ordered by number of regular season fantasy ‘Wins’.  Each statistical category is totaled up through September 22nd.  The symbols next to each team abbreviation indicate teams that made the playoffs (*** for division winners and # for Wildcards).  While team CR had the same number of wins as BBK, CR advanced by the tiebreaker.

How Else Can We Look At The Standings?

When I look to independently evaluate my team performance, I try to find ways in which I can evaluate everyone on a level playing field.  That means finding ways around scheduling issues and luck due to bad weeks at the wrong time, or great weeks when the opponent was just a bit better.  This year, I decided to get an idea of the distribution of each category by standardizing the totals of each of the teams in my league to find where I stand on that distribution.

To do this, I first calculated the Mean and Standard Deviation of each category total in the League Standings. 

I followed up by calculating a Z-Score for each team in each category, or the number of standard deviations each team category total is away from the mean. Using these, we can look at individual categories, or we can find a way to rank all of the teams based on their set of Z-scores. 

The advantage to the standardized score is that the values in each statistical category are easily comparable.  In addition, I was able to develop a Composite Z-Score by averaging each team’s 16 categorical Z-Scores.  This puts the rankings on a continuum, rather than a clunky Rank scale that results in the 100R = 1R problem in discrete scoring. 

All of these metrics are very easy to calculate and Microsoft Excel will do it automatically with its functions tab.

So were there any teams that were extremely unlucky?  Below are the rankings based on the Composite Z:

Composite Z shows us that BBK ran into quite a buzz saw, and not just due to the playoff tiebreaker.  This team had a tough schedule and/or bad luck that cost the owner about 7 places in the standings. While the team put up fantastic overall production, the end result was a 10th Ranked record and his team missing the playoffs despite fielding a roster that ranked 3rd on the Composite Z scale.  BB and KI got the largest boost in wins thanks to their weak divisions, though BB looks to be a playoff caliber team with or without the weak schedule.

What Can We Take From This?

First, we have to remember the caveats of the metric. 

Because this is a league that allows daily lineup changes, owners could be maximizing wins in the short run, rather than trying to gain overall category totals (as they would be in a standard rotisserie league).

Also, an extreme individual category Z-Score could result in overestimation of the Composite Z.  If there is a worry of this problem, one could use Median Z rather than Mean Z. 

Finally, there could be a problem with the ‘normality’ assumption of the distribution of the statistical categories.  So while this is another nice metric to evaluate the caliber of roster building, it’s not an end-all statistic.  

We can observe that the Orange Division was by far the most competitive, with the Blue and Green Divisions being much less so.  The Red Division is interesting.  This is an extremely top-heavy division, as ML, FF and TDG appear to all be playoff-caliber teams.  Normally, this would hurt these teams’ records, but thanks to the terrible IF and SS, the top 3 teams were able to scrounge up enough victories to make the playoffs.  

Individually, if I were BBK, I would consider my season very successful and try to take a similar approach to roster building the following year.  Perhaps trying to find strategic moves that would maximize my wins at the margin would be beneficial in the future.  On the other hand, if I were the owner of KI, I would have to think about where my team needs to improve.  I could now understand that I cannot rely on running into this kind of luck next season.  

All in all, the Composite Z seems like a useful metric to get a slightly different view of league standings.  The best part is it’s infinitely easier to compute than some sort of probability model and allows for a dynamic look at categorical standings we may not have otherwise been able to see.