Adjusting to the New Reality

Adjusting to the New Reality

The level of offense in baseball has been dropping for some time now. In the 1980s and into the early 1990s, teams scored around 4.3 runs/game (with the exception of 1987, when offense jumped up to 4.7 runs/game for one year, then went right back down in 1988). Offense started to rise in 1993 and first jumped over 5 runs/game in 1996. Run-scoring peaked at 5.1 runs/game in 2000, then leveled off to around 4.8 runs/game through 2007. Since 2008, offense has gone down steadily, with 2014 seeing an average of 4.1 runs/game. You have to go back to 1981 to find fewer runs per game in baseball (4.0 runs/game).

This has implications in the world of fantasy baseball. Consider the table below that shows the ERA in Major League Baseball by year, going back to 2001:

YEAR	ERA
2001	4.42
2002	4.28
2003	4.40
2004	4.46
2005	4.29
2006	4.53
2007	4.47
2008	4.32
2009	4.32
2010	4.08
2011	3.94
2012	4.01
2013	3.87
2014	3.74

 

Some would point to PED testing for the lower level of offense, some would blame a bigger strike zone, some would peg it on the increasing number of relievers throwing 95+ for an inning or two. Whatever the reason, this is the new reality and sometimes it can be hard to adjust to new realities.

Let’s look at the numbers shown above in more detail.

Over the stretch of years from 2001 to 2009, MLB had an ERA of 4.39. Over the three-year stretch from 2010 to 2012, ERA dropped to 4.01. The last two years have seen big drops each year, from 4.01 to 3.87, to 3.74.

This has repercussions in fantasy baseball. With ERA dropping quickly, we need to reevaluate the pitchers we take on draft day and during the season.

Let’s go back to 2009, when MLB had an ERA of 4.32. The top 60 starting pitchers in ERA (minimum of 160 IP) combined for an ERA of 3.54. The median ERA for this top 60 was 3.77. There were 11 pitchers with an ERA under 3.00.

Fast forward to 2014. Last year, MLB had an ERA of 3.74. The top 60 starting pitchers in ERA (minimum of 160 IP) combined for an ERA of 3.14. The median ERA for this group was 3.33. There were 22 pitchers with an ERA under 3.00.

	2009	2014
ERA in MLB	4.32	3.74
ERA of Top 60	3.54	3.13
Median ERA of Top 60	3.77	3.33
Pitchers under 3.00	11	22

 

In 2009, the median guy in the top 60 was someone like John Danks (3.77) or Jarrod Washburn (3.78). Last year, the median guys in the top 60 were Jose Quintana (3.32) and Chris Archer (3.33). [Caveat: I know ERA isn’t the only way to judge a pitcher in fantasy baseball. I’m keeping it simple.]

Six years ago, when scouring the waiver wire, that pitcher with a 4.00 ERA was a potential pick-up. These days, you don’t want to look at that guy, he’ll just hurt your team. This may seem obvious, but it really is a change in mindset when you’re looking to improve your team. What we once thought was good is no longer good.

One of the side effects of a big drop in the run environment is the difficulty for projection systems to keep up. If we go back to the 2010 season, we can see a stark example. If a pitcher had league average ERAs in 2007 (4.47), 2008 (4.32), and 2009 (4.32), we could do a simple 3/2/1 weighted average for his three seasons and project an ERA of 4.35 for 2010. League-wide, though, ERA dropped from 4.32 in 2009 to 4.08 in 2010. Most projection systems will project ERAs that will be in line with the previous few seasons’ run environment. In this case, the projections will be well above what the actual ERAs were for the 2010 season (unless a projection system can anticipate such a drop in offense).

Let’s do the same for more recent seasons. If we take a pitcher with league average ERAs in 2011 (3.94), 2012 (4.01), and 2013 (3.87), and do a simple 3/2/1 weighted average, we get a 2014 projection of 3.93. The actual ERA in MLB in 2014 was 3.74, so pitchers as a group are going to be forecast with ERAs around 0.20 higher because the drop in offense was so drastic.

With this in mind, I looked at last year’s projections from four systems: Steamer, ZiPS, Davenport, and Oliver. I looked at all pitchers who were projected by each of the four systems who pitched 30 or more innings in 2014. There were 326 pitchers in this group and they finished 2014 with a combined ERA of 3.58. You can see how each of the projection systems forecast these players prior to the 2014 season:

2014 SEASON
Actual ERA	3.58
Davenport projection	3.76
Oliver projection	3.81
ZiPS projection	3.90
Steamer projection	3.91

 

When looking at the data, what you shouldn’t do is say that Davenport had the best projections. What is true is that Davenport best anticipated the run environment. Looking at the table, it would be easy to assume that Davenport and Oliver had the best projections, as they were closest to the actual ERA of this group of pitchers. In reality, if you are trying to assess which system better projected individual players, you would first want to adjust them all to the actual run environment, then compare the differences between projected ERA and actual ERA for individual pitchers.

In the case of the 326 pitchers used above, the table below shows the average absolute difference in actual ERA and projected ERA for each individual pitcher, using projections adjusted to the run environment of this group of pitchers.

Adjusted Projections
System	AvgAbsDiff
Steamer	0.85
Davenport	0.86
Oliver	0.88
ZiPS	0.90

 

Looking at it this way, it’s easy to see that the different projection systems were very close on this group of 326 pitchers and Davenport and Oliver are in the middle of the pack, with Steamer moving from the bottom to the top.

What does this mean for 2015? If you’re the type of fantasy baseball player who likes to create your own projections by combining projections from other sources, you will first want to know what level of offense those projections are expecting (ERA in this example). If you think 2015 will be much like 2014 (3.74 league-wide ERA) but the projections expect an ERA much higher or lower, you should adjust all pitchers by the amount the projections are high or low. With these new adjusted projections, you can now combine your projections.

As an example, I took those same 326 pitchers from above and compared their actual combined ERA from 2014 to their 2015 Steamer projections. This group of pitchers had a combined ERA of 3.58 in 2014. Steamer is projecting them to have a 3.84 ERA in 2015. The difference is 0.26 in ERA. I don’t know the run environment Steamer is basing their projections on, but this would suggest that it’s higher than what we saw in 2014.

Based on the disclaimer that accompanies each team’s ZiPS projections, we know that ZiPS is projecting based on the AL having an ERA of 3.93 and the NL having an ERA of 3.75. This would be a slight increase from the 2014 season (AL: 3.82 ERA, NL: 3.66 ERA) and is, essentially, a 3/2/1 weighted average from 2012, 2013, and 2014.

I looked at the starting rotations for the five teams that we have ZiPS projections for so far. There are 25 pitchers and they are projected by ZiPS to pitch 3985 innings with a 3.73 ERA. These same 25 pitchers are projected by Steamer to pitcher 4039 innings with a 3.98 ERA. Steamer is high by 0.25. Steamer projects higher ERAs for 23 of these 25 pitchers. This is a small sample of just 25 pitchers, but it would appear that you will want to adjust the Steamer pitching projections down if you do any sort of combining of projections in your fantasy baseball prep.

In addition, if you’re in a keeper league and have access to last year’s data for your league, you may want to project your keepers and potential additions for 2015 and compare your team projections to last year’s stat categories. This way, you will have an idea of how competitive your team will be. For example, I’m in an 18-team, 25-man roster league. We have nine starters on offense, four starting pitchers, and two relievers in our active lineups, and a 10-man bench that can be made up of players from any position. Teams in this league averaged around 1000 innings last season, so when I create projections, I can plug in the stats for my keepers and potential additions to see how my team looks for the upcoming season. In order to compare my projected 2015 team to 2014 stat categories, I want my projections to be adjusted to the level of offense of 2014 (in this case, ERA).

Offense in baseball has been dropping for a few years now. Successful fantasy players will have to adjust to this new reality when doing their pre-season prep work, on draft day, and when adding players from the waiver wire.