There have been a number of improvements behind the scenes on the daily projections (available now!) that I wanted to share before I forget them all. Here it goes:

  • As of last week, all Hitter projected Plate Appearances are tied to last 30 day averages (instead of pre-season projections). Adjustments are made based on whether the opposing pitcher is RHP vs LHP. This is a slight improvement from May-September hitter projections in 2014 as they did not reflect the handedness of the pitcher (e.g., Rajai Davis more likely to hit at top of lineup against LHP than RHP)
  • As of last week, all Pitcher projected innings are tied to last 30 day ‘Batters Faced’ averages instead of pre-season projected IP. This is an improvement from last year where the IP projections did not factor in recent starts (the pitcher streakiness analysis I did in the offseason drilled home the value of making IP calculations more dynamic).
  • Starting today, a hitter’s estimated percent chance of being in the starting lineup (%St) has been updated to be more dynamic. Any player who has started at least 4 games in the last 7 will automatically have a floor of 65%. This helps both the utility of Hittertron as well as improve the Streamonator estimates since it leverages this stat to determine opponent’s offensive strength.
  • Starting today, all Hitter daily projections now incorporate opponent bullpen strength. The ratio of Plate Appearances against the starter vs the bullpen has also been adjusted slightly. Previously, PAs credited to bullpen were estimated as against league average pitching.
    • Bullpen projections are driven based on last 30 day team usage as well as the projected starter IP. Injured and demoted pitchers are not included. For today’s projections, relievers are removed if they pitched 2 days in a row or pitched past a certain threshold of batters the prior day.
    • In a test using May 4th-May 13th projections, the new results led to an average decrease of negative $2.68 across all players projected at 70+% to start ($10.46 to $7.78). The negative shift is driven by the fact that relievers have better rates than the average pitcher inning.
  • In mid-April, we added the following stats in the Player Page Game Logs:
    • Hitters – Added ‘BatOrder’ (where a batter hit in the lineup), Opposing Pitcher, and Opposing Pitcher Hand (L/R)
    • Starting Pitchers – Added Batters Faced
    • Relievers – Added ‘Batters Faced’ and ‘IP Start’ (the inning a pitcher started).

There are several other enhancements we are hoping to get out this year. The one that is experiencing delays right now is customization of league settings for Player Raters/Streamonator/Hittertron. I will post another update in early June.

 

 

  1. Thanks, Rudy

  2. Nation says:
    (link)

    Great work as always rudy!

  3. Me says:
    (link)

    you continue to be the lifeblood of Razzball. you deserve more credit man :)

  4. Clint says:
    (link)

    Amazing stuff as always Rudy. I feel like Peter Griffin when I read it, though.

      • Clint says:
        (link)

        @Rudy Gamble:
        Question for you: I know the lineup tools help you choose between a middling two start guy and a closer on a weekly basis but can it also be done to optimize a head to head points league lineup on an overall basis? Or have you done any studies to see what the optimal pitching lineup would be in such a league w/2 SP’s, 2 RP spots, and 4 P spots?

        Reason I ask is that this is my 2nd year in such a league & I’m still getting a handle on whether maximizing starts at the expense of any bench bats using RP/SP guys or rostering any closers also maximizes your points potential per week. I see some guys not even rostering a catcher to keep a SP on their rosters while others have several bench bats and/or every closer they can get their hands on. I avoided closers last year, used my 6 moves/week for SP pickups at a streamer spot, and kept all bench spots for pitchers & won the 8 team keeper league. Grey advocated keeping 2 closers for the 2 RP spots, 1 streamer, and the rest pitchers this year & I’m 4-0 so far but not as high in points as I was last season. Wondering if you have any tools or evidence to support that or a specific configuration for head to head formats?

  5. Pine Tar Incident says:
    (link)

    Thanks for the update Rudy. I really hope customization is able to happen! Please get the Stream-o-Nator to stop forcing me to start Nate Karns :)

    • ha. not sure customization will help if SON loves him.

  6. Steve Stevenson says:
    (link)

    That’s dope as hell, Rudy. I’ll buy you a Mexican martini and a taco next time I’m in Austin.

  7. Fastpitch says:
    (link)

    This is great for us stat/number guys. After all fantasy baseball is like the stock market. Each player is a stock, pick the right ones and you are rich or in 1st place. Thank you for informing us of the updates.

  8. Fastpitch says:
    (link)

    Rudy, I have a question. For example L. Martin has a wrist injury and is day to day. Is the onus on me as a manager to know this? When does hitter tron or son update as the injuries happen? I’m thinking what we see is if the player is healthy.

    • @Fastpitch: The day after a player goes on the DL (7, 15, or 60) he is removed. If a player goes on the DL or comes off it, I often catch it and re-run the results. Day-to-day, though, I can’t do anything about. Hoping to set something up soon that takes BaseballPress lineup data and removes any non-starters for today’s games.

  9. Jack says:
    (link)

    Literally better every day!

    Wondering if you could make an option available for Hitter-tron. Make an option which would allow the user to display all available days (like Stream-o-Nator). This would be useful when planning for H2H matchup with Daily changes.

    • @Jack: Thanks. I used to have that but the table size was way too big. It was sooo slow. Easier to do with pitchers since there are two per game vs 20 batters.

  10. Uncdrew says:
    (link)

    Is a negative decrease an increase. Is a false negative a positive. Inflammable means flammable? Inconceivable. I love this country.

  11. BigUps Cubs says:
    (link)

    Just made my first deal of the season. Carlos Gonzales for Craig Kimbrel. Have OF depth with Carlos Gomez, Justin Upton, Starling Marte, Lorenzo Cain and Jake Marisnick. Needed RP help badly with Ottavino going down and only having Wade Davis and Cishek rostered.

    Thoughts?

  12. Chris says:
    (link)

    Excellent stuff, thank you very much Rudy. All the changes sound great and am glad you took the time to highlight them here!

  13. Hal Incandenza says:
    (link)

    Just wanted to echo the thanks, Rudy! More customization sounds awesome, so best of luck there.

  14. biffrenylds says:
    (link)

    Rudy, the best just got better! One question: have you ever played around with weather effects? dailybaseballdata.com posts wind speed/direction as well as air density for each game. Could that provide extra benefit beyond simple park effect?

    • @biffrenylds: Hey Biff – definitely familiar with dailybaseballdata as we link to their hitter vs pitcher matchups and they link to our player pages. I haven’t played around with weather – to factor it in, I’d need the weather data for a number of past games to see if it helps improve the estimates. My guess is that – aside from rainouts – it’s probably a non-factor.

  15. Kevin S. says:
    (link)

    Good stuff, Rudy. Appreciate all the hard work. I use the Razzball projections daily.

  16. Scott says:
    (link)

    Thanks!

  17. goodfold2 says:
    (link)

    big change from average hitting values being replaced by specific team bullpen values. I figured this was already done before. very good update there.

Comments are closed.