Ease factors

phillydilly  4/1/2013 11:23 AM

Had a question about EASE Rankings.

I'm looking at todays SP. I would think Nolasco, Worley, and Chacin are the 3 worst pitchers starting today.

They are playing Was, Det, and Mil respectively.

Using Draft Days Scoring system, i have the following without EASE

Braun: 19.62

Miggy: 19.22

Bryce: 15.6

If I apply EASE, I get

Braun: 18.62

Miggy: 17.68

Bryce: 15.27

So unless I'm doing something wrong, their point values are going down going against worse pitching.

Can you clarify how EASE is calculated?

ken_admin  4/1/2013 1:05 PM

Yes, this doesn't make sense, and look into it further, here's what's happening:

We currently compare the hitters average to what the pitcher allows.  We adjust the hitter up/down based on the difference.  What I've found is that most starters tend to allow less than the top hitters produce so it tended to move all of the players stats down. The good news is the drop due to this occurs more for bad pitchers than good pitchers so the overall rankings still show the correct effect for ease.

To correct this, we now find what the pitcher allows and compare it to what the average pitcher allows, and then adjust the hitters based on this.

I will apply this fix today. Thanks for pointing it out.

bcavers  4/1/2013 2:01 PM

you also plan to add park factors correct?

ken_admin  4/1/2013 2:15 PM

Yes, we will be adding that soon.

Rapierman   4/3/2013 11:26 PM

Do Ease Ratings take into account certain situations that might be instantaneous (i.e., batter-vs-pitcher, lefty-vs-righty, etc.)?

ken_admin  4/4/2013 6:59 AM

We're working on adding more factors, but currently we take into account the strength of the pitchers for the hitter projections.  The following are also planned: strength of offense for pitchers, L/R matchup, and park factors.  We are now showing a Matchup column with the opposing pitcher, their L/R along with the hitter L/R.

#28675  4/9/2013 2:20 PM

I'm very interested to know when Ease takes into account things like:

L/R splits

PvB splits

Park factors

ken_admin  4/9/2013 2:45 PM

We're done with the Remaining Projections and will be making changes to the daily next.  I would expect these by the end of the week at the latest.  Sorry for the delay.

bowiac  4/9/2013 5:44 PM

Can I inquire how are these issues going to be handled? I ask because there's no consensus as to how to deal with this. For instance, with L/R splits, there's research suggesting all righties have the same 13% platoon split, and that deviation from that is basically random (exarcerbated by sample size). But then again, others believe some hitters have large/small platoon splits, and need specific adjustments.

The same goes for park adjustments. Any time you divide the sample by two, you risk small sample size flukes.

Are the L/R and park adjustments going to be generic, or individualized? I'd be very interested in applying generic adjustments (13% platoon split, or 25% extra run scoring at Coors) but less so with specific ones that'll cause bizarre results due to small sample sizes.

Can you shed some light as to how this will be done?

liam4  4/10/2013 6:43 AM

Ken,

Can you please clarify for me what ease currently covers, for hitters?  Is it strength of the pitcher and strength of the pitcher's team (per Ease rankings), or just the former?  Any other aspects? 

Additionally, when I run daily projections but do not adjust for Ease, the results often differ from annual per game projections; one player might be ranked ahead of another in one, behind in the other.  What factor is that taking into account? 

Thanks!

ken_admin  4/10/2013 10:21 AM

The bulk of the adjustments will be made on the general strength of the opposing hitters/pitchers.  We have underlying stat inputs that generate the end projection, and those will be adjusted and then the projections will be generated.

For L/R, we will use a general adjustment rather than individual hitters vs pitcher history.  Park adjustments will also be general adjustments.

For all factors, we will allow users to see the effects of each and optionally turn each on/off.  The planned factors are:

Opponent Strength

L/R Adjustments

Park Adjustments

ken_admin  4/10/2013 10:28 AM

Currently, we're only considering the strength of the opposing pitcher when adjusting hitters but are currently adding more (opposing hitters, L/R, park).

The daily results without ease may differ from the general long-term projections due to the value calculations.  For long-term, we calculate the top players and use those to determine the z-scores for each category. When doing daily, we do the same, but the resulting pool of top players will be different causing small differences from the long-term projections.  Plus the daily projections use actual lineups while the long-term use estimated lineups.

#28675  4/10/2013 10:44 AM

Thanks k. That seems very reasonable.

liam4  4/10/2013 11:19 AM

Thanks Ken.

One question regarding hitters facing a Colorado pitcher; is there currently a distortion to take into account when the game is not held in Colorado?

ken_admin  4/10/2013 11:21 AM

Currently, park factors are not considered, but they will be.

ken_admin  4/14/2013 10:42 AM

The factors are still on their way, and you can see the following on our preview site:

Opposing Pitchers

L/R Adjustments

http://preview1.baseballmonster.com/DailyProjections.aspx

Park factors and Opposing Hitters are coming next, and then we'll deploy all factors to the website.

#28675  4/15/2013 12:29 PM

Solid. Thanks for the update! Looking forward to the park factors, particularly.

Rapierman   4/15/2013 10:51 PM

I'd love that.  One has to consider:

• Grass vs. Artificial Turf
• Indoors vs. Outdoors
• Dimensions of the stadium
• Home Plate Umpire Tendencies

#28675  4/16/2013 8:44 AM

Park factors should capture a lot of that, rapierman.

Umpire effect... should be factored in as generic "home field advantage" which is mostly umpire (if not all).

Rapierman   4/16/2013 5:36 PM

Actually, "Home Plate" would also take into count how the govern the strike zone (high strike, low strike, wide zone, narrow zone)