Don't be shellfish...Share on FacebookTweet about this on TwitterShare on Google+

There are three things I look for in my fantasy baseball projections – cheap, early, and accurate.

So I’m excited to announce that we now have 2013 fantasy baseball projections available on Razzball for free – courtesy of Steamer projections – before both the groundhog and other projection systems have popped their heads above ground.  Steamer moved up their number crunching a couple weeks from last year and we helped by providing playing time estimates (up to and including the Justin Upton trade).

As for accuracy, Steamer projections – created in 2008 and maintained and operated by Jared Cross, Dash Davidson, and Peter Rosenbloom – tested as one of the best (if not the best) in my review of 2012 baseball projection systems.

We will update these projections periodically throughout Feb/March as MLB depth charts stabilize and Scott Boras finds a hefty contract for his last overpriced client.  The projections can be accessed via the ‘Stats’ menu or by clicking the below links.  (Note:  Grey’s projections will be available in the same menu very soon.)

In addition, the first release of our $ Values + Point Shares are available.  They cover 5×5 leagues with 10, 12, and 15 teams using the ESPN and Yahoo roster formats (AL/NL-only $ values will be available later in February)  These leverage the Steamer projections as well as CAIRO projections which performed well in the aforementioned 2012 projections test.   These can be accessed via the ‘Player Rater’ menu or by clicking the below links and will be updated periodically throughout Feb/March.

For those too lazy to click, here is a sneak peak at the top 10 players according to my $ values for ESPN 5×5 12 team leagues:

Rank Name Team Pos Point Shares $ Value
1 Mike Trout LAA OF 8.3 $42.0
2 Ryan Braun MIL OF 7.7 $39.7
3 Miguel Cabrera DET 3B 7.4 $38.4
4 Albert Pujols LAA 1B 5.8 $32.4
5 Giancarlo Stanton MIA OF 5.2 $30.1
6 Clayton Kershaw LAD SP 6.0 $29.9
7 Robinson Cano NYA 2B 4.9 $29.2
8 Justin Verlander DET SP 5.5 $28.0
9 Justin Upton ATL OF 4.1 $26.1
10 Matt Kemp LAD OF 4.1 $25.9
  1. Jim says:
    (link)

    Adam Eaton’s projections should be updated.

    • @Jim: His increased playing time is reflected in the Point Shares/$ Values and will be addressed in the projections for the next update (by mid-Feb).

      Here is the underlying projections behind his Point Shares/$ Value:
      410/66/7/40/22/0.293

      If you believe Eaton will play 150 games, you could multiply everything but the average by 1.25. For now, I think that Parra will be rotated in more often than the average 4th OF and, thus, all the ARI OFs are projected a little lower for playing time.

        • zk says:
          (link)

          @Jim: Please. These are free projections. Say please. EVEN ON THE INTERNET YOU SHOULD SAY PLEASE.

  2. popups and popovers says:
    (link)

    Is there any way to adjust for single league formats? I would imagine that the average player against which the point shares are measured would be different when you remove half the player pool (offense stats are probably inflated in the AL, etc.)

    • @popups and popovers: Well, you can look at the 2012 Player rater data (razzball.com/playerrater) and compare the mixed league vs. single league values. It’s difficult because single leagues will have varying position scarcity/depth vs. mixed league. A solid example is last year’s 1B where the AL had very good depth (Pujols, Miggy, Fielder, Teixeira, A-Gonz) while the NL had Votto and….honestly, I think Freeman was #2 pre-season 1B….

      I’ll get single league $ values up there sooner vs. later. Just kind of fruitless when there are still free agents mulling about.

  3. Tyler says:
    (link)

    Rudy, I always thought you were the best. Well, you or Jared…..I could never decide. Now i don’t have to! Should have known it was a matter of time before y’all joined forces.

    • Thanks! I’m psyched we were able to join forces too. Jared is WAY better than me on the hard-core statistics stuff. I have him (for now) in Excel skills though :)

  4. royce! says:
    (link)

    Nice, thanks Rudy! Lots of value picks in here, if the projections pan out.

    How do you feel about the 177 IP projection for Strasburg? Is it reasonable to just add ~ 20 IP to last year’s total? If he reaches 200 IP with the projected peripherals, seems he would end the season the top SP.

    • This is the first year that I’m actually making playing time projections from scratch (I used to use a couple sources, average them together, and tweak here and there).

      For IP, I focused first on # of starts and then, secondly, on IP/GS from last year.

      I gave Strasburg 28 GS which assumes he has a 15-day DL trip or gets skipped a couple times. I don’t feel comfortable projecting any starter with 32 GS (162/5) unless they’ve done it at least once before.

      Strasburg was the one pitcher where I increases his IP/GS from last year because he averaged only 90 pitchers per start. So I upped that to 100 pitches. That increases his IP/GS from last year’s ~5.5 IP to an estimated 6.3 IP. To project him at 200 IP, he would need to not miss a start or he’s going to have to pitch over 7 IP/GS which would mean he’d have to be a lot more efficient with his pitchers than last year.

      That said, if I project him at 200 IP through one of those two routes, he jumps to a virtual dead heat with Kershaw as best projected SP.

      • Tyler says:
        (link)

        @Rudy Gamble: Just out of curiosity, why did you decide to do scratch PT projections this year?

        Also, is there a page that describes how we can calculate PS’s for our own custom leagues?

        • @Tyler: I answered this in another post but basically the sources that I think do a solid job at Playing Time don’t report until early March. The only source I’ve seen available so far is RotoChamp and they tested just so-so in 2012.

          • Tyler says:
            (link)

            @Rudy Gamble: Makes perfect sense, thanks.

  5. Jeff says:
    (link)

    Awesome! Thanks Rudy. Do I recall correctly that you also do a version where OBP replaces average?

    • @Jeff: I am going to have that in place for the in-season player rater but I’m a little more limited in what I can do for pre-season values. Will take it into advisement.

      • SwaggerJackers says:
        (link)

        @Rags: 3rd’d!

        • @SwaggerJackers: Ok, it’s done. Added in hitter values for 5×5 (OBP replacing AVG) as well as 6×6 (adding OBP) and 6×6 (adding OPS). All based on 12 Team ESPN leagues.

  6. Joe says:
    (link)

    This is probably the best tool I’ve ever come across!!! The values appear to be very close to what the actual auction will be more than I have ever seen before. In the past, I have waited for “expert” leagues to hold their auctions, and, from that, figured the inflated values for my leagues.

    Quick (and possibly stupid) question, though…what accounts for the differences in value between the ESPN and Yahoo leagues?

    • @Joe: Thanks! The difference in ESPN and Yahoo leagues comes down to format and ratios. ESPN has 13 hitters and 9 pitchers. Yahoo has 10 hitters and 8 pitchers. But if you look at year-end totals for Yahoo, you’ll find that it’s more like 9 pitchers as the IP per team is about equal to ESPN leagues.

      Because Yahoo has less players (19* vs. 22), the average player is worth more $ in Yahoo.

      Because the ratio of hitters to pitchers is smaller in Yahoo, it makes hitters more valuable in Yahoo than in ESPN (e.g., MIggy Cabrera’s stats represent 1/10th of our totals but only 1/13th in ESPN).

      I’ve found that the resulting differences are significant enough that I create different player raters for each (the preseason survey helps confirm the popularity of both roster formats as well).

  7. Ferga says:
    (link)

    Would you happen to know anything about the lastplayerpicked guide with these projections? I like that I can use that customizable guide with steamer’s full projections for my obp/slg and holds league.

    Would love it if your site could do something like that.

    • Rags says:
      (link)

      @Ferga: Seconded. Lastplayerpicked’s site seems to be down. Don’t know what happened to the guy, but his updates were getting spotty last year. Maybe he lost interest?

      I ran and reran numbers on that site over and over (not just for draft but to help with midseason trade values too) and would love it if Razzball did something similar…

        • Rags says:
          (link)

          @Tyler:

          :D !!!

          hooray!!!

      • We had an in-season player rater that updated near-daily last year. You can see end of year results under the “Player Rater” menu.

    • @Ferga: I am working on supporting a couple of extra formats besides 5×5 for in-season projections. I don’t think I’ll have that in place for the pre-season projections though.

      We have talked about developing a tool for custom league formats. It’s on the consideration list for next year’s enhancements.

    • @Ferga: FYI, Holds projections now available under Pitcher Projections. Will add them soon into the $ Values/Point Shares.

  8. Jake says:
    (link)

    Rudy, one of the things I noticed from your 2012 projections analysis was that Steamer rated as one of the best in predicting playing time, so I have been eagerfor their 2013 playing time projections. However, here you say that you provided the playing time projections rather than Steamer.

    Now, I don’t mean to denigrate your predictions, but I’m a little confused: for these projections, are we working from Steamer’s battle-tested methods or from your own unknown methods of calculating playing time?

    Does my question make sense or am I completely missing the point?

    Thanks!

    • It’s a great question. The folks at Steamer decided to forgo creating their own playing time estimates this year to focus on other aspects.

      We considered using other sources like BaseballProspectus, Fantistics and FanGraphs but they either aren’t ready until early March or are too shallow (FanGraphs Fan Projections). There is another reason why we took ownership of creating playing time estimates which will become clearer as we announce all our upcoming enhancements for 2013.

      So these are not battle-tested playing time projections but I will test them at the end of the year using various snapshots (e.g., how did mid-Jan compare to mid-Feb compare to mid-March, etc.

      If you have a preferred source for playing time, you can just download the projects and re-weight the counting stats.

      • We’re very excited to be partnering with Razzball.

        @Rudy Gamble: My impression is that Rudy is doing playing time the same way we did it for Steamer but doing it better.

      • Jake says:
        (link)

        @Rudy Gamble: Thank you! I’m looking forward to seeing what else you guys have coming.

  9. Rags says:
    (link)

    are you going to do 2 catcher leagues when you do AL-/NL-only?

    Also, I see the 160/100 and 180/80 budget splits and I know this is a common option, but from a mathematical standpoint, shouldn’t ideal proportions be calculable?

    I know there’s the arguments about different replacement levels for hitters and pitchers, injury concerns and projection reliability. But isn’t there some way to test which budget ratios get the best results?

    It’s something I’ve been curious about for a while, but I don’t think I have the math or statistics chops to figure it out myself.

    • Rags says:
      (link)

      @Rags: @Rags: Also, sweet! Thanks! Love this stuff!

    • Yes, AL/NL only will be based on the 2 catcher format only.

      The ideal budget split depends on the drafter’s POV. I’ve always felt that there was a drafting bias towards hitters. The 2012 projections test I did showed that hitter projections are, on average, more reliable than pitcher projections. It seems that Steamer’s pitching projections have helped to reduce that gap. To account for the remaining gap, I did tweak the $ estimates so that $160/$100 is probably closer to $163/$97. This is ONLY for pre-season projections though as the in-season player rater will remain at $160/$100.

      I provide the $180/$80 because it is a more accurate reflection of how most people draft – especially in snake drafts. Since most people draft this way, I tend to space out my pitching picks to where I get the best ‘bang for the buck’ and/or where I feel I’m paying the lowest premium possible for hitters.

      Hope that helps…

      • Rags says:
        (link)

        @Rudy Gamble: How did you get the 163/97, though? And does that change significantly from format to format? Most of the time, I just get the impression that people are spitballing these numbers…

        • I screwed up the explanation a bit. A perfectly proportional allotment would be $156/$104 (13/22*$260 / 9/22*260). Hitter projections using Steamer 2012 were about 5% more predictive of RCL team results than the Pitcher Projections. So I adjusted the hitter pool up by 2.5% and the pitcher pool down by 2.5% which nets out to $160/$100.

          • Rags says:
            (link)

            @Rudy Gamble: So that’s just based on the number of hitters and pitchers for each team?

            Should the relative replacement level of hitters vs. that of pitchers have some effect though?

            • That’s a tough call to determine the relative quality of replacement values between hitters and pitchers. I find replacements values hazy to being with given that deep positions get snapped up disproportionately for UTIL and people routinely have pitchers on the bench (and thus not eligible for replacements). That’s why Point Shares focuses on ‘average value’ as it’s more straightforward (at least to me) to determine the universe of drafted players.

              The best arguments I’ve considered (that I haven’t accounted for like differing projection accuracy) are:
              1) Streaming SPs (squeeze okay to good results from undesirable pitchers in good matchups)
              2) Market inefficiency when it comes to Middle Relievers (cheap, fungible ERA/WHIP/K helpers)

              I think #1 is a compelling factor with the massive caveat that the strategy’s effectiveness goes down immensely if you have multiple owners doing it (especially in RCL where everyone knows about the Stream-o-nator). We’re aiming to launch something similar for hitters that might balance the hitter/pitcher dynamic as well (e.g., identifying FAs with favorable matchups and/or rostered hitters who have unfavorable matchups).

              • Rags says:
                (link)

                @Rudy Gamble: I’m actually in a league with transaction fees and no bench (except for injuries) so streaming is almost non-existent in my league. Your point is taken though.

                On the one hand, it would seem to me, given my format that I should spend more on pitchers since you have to make what you have count. On the other, I feel like more usable starting pitchers are always available on the wire than any other position, so I should spend more on hitters.

                As for middle relievers, being in an NL-only keeper league, I basically have to take other rating systems, and completely throw out reliever numbers. I’ve never seen a rating system that accurately described closer prices in my league. People are paying so much for closers, there’s a creep where people are starting to pay more for good 8th inning guys in the hopes of an injury or something.

                It’s a weird format I’m in, but I feel like it should actually make the math easier. I just feel like a comprehensive value system is always just out of reach though.

                In the meantime, thank you guys for something that’s probably close enough…

                • I’m not convinced that there are more good pitchers vs. hitters on the waiver wire in the first month or two of the season – especially in shallower leagues. I was able to grab Adam LaRoche and Kyle Seager in the RCL during the first month of the season. But after that first month or two, it does seem like playing time solidifies and it’s easier to find pitchers vs. hitters.

                  I agree that most single-league formats sometimes overbid on non-elite relievers. It’s a scarcity play. I find the best bargains are either paying $20 for a great reliever (Kimbrel) or $12 for a high-K reliever that is on a bad team (e.g., Grilli). That said, take a look at last year’s NL 12 Team player rater (http://razzball.com/playerrater-nl12team/) and type RP in the filter. I had Kimbrel and Chapman worth around $29! I actually valued Kimbrel at $27 in the pre-season. I had Chapman higher than anyone else at $9…and that was when Marshall was set to be the closer. Love me reliever K’s!

                  • Rags says:
                    (link)

                    @Rudy Gamble: Yeah, it doesn’t help that everyone’s speculating on saves in a keeper league, so inflation is through the roof, too. All of the closers in our league are either kept in the 8-10 range because they were picked up before they were closing, or they’re in the mid- to high-twenties like you valued.

      • birrrdy! says:
        (link)

        @Rudy Gamble: ugh, here’s a vote for 1 C only when you do your NL only… A huge chunk of us seem to play NL or AL only these days, and the 2 C format is starting to wane from my anecdotal observations… I know you did data/surveys on this, so will obviously defer, but thats my 2 cents…

        • Here are some survey results:
          – Roughly 25% of respondents are playing 1+ AL/NL only leagues (whereas mixed leagues close to 100%).
          – There’s a 50/50 split in AL/NL only between 1 and 2 catchers with the 1 catcher leagues split again by ESPN vs. Yahoo format.
          – There’s close to an equal split between 10 and 12 team leagues with other leagues sizes much lower.

          Verdict:
          – At this point, doing 10/12 team 2 catcher / ESPN/NFBC format would match the most people.
          – Can consider more in the future but it would be fighting against resources for more mixed league formats.

  10. Tyler says:
    (link)

    Rudy – not trying to spam this question, but wasn’t sure if you saw it; is there a page that explains the entire methodology of the PS calculations (such that I could calculate it for a league with custom categories and positions)? I’ve seen this page (http://razzball.com/2010-razzball-projected-point-shares-ver-1-12-team-mlb/), but not sure if there’s another I’m missing…

    • @Tyler: I don’t have a page available that lays out the entire methodology. This is out of equal parts laziness and protectiveness (word?). There’s a lot of work to set it up from scratch. Each year I’m getting better and better about scaling it so I can do more formats. We’re moving off Excel to MySQL this season which will hopefully set the groundwork to support more formats (and possibly a custom league tool).

      Sorry I can’t be of more help….

      • Tyler says:
        (link)

        @Rudy Gamble: Gotcha – I’m not looking a gift horse in the mouth. Thanks much for all your work.

  11. pythons007 says:
    (link)

    In my auction players don’t come anywhere close to these projections. This is the first time I’ve seen, this and I’m curious why players are worth negative amounts. The highest anyone has gone in my league was mid 40s and I know every league is different. So how should we take these salary projections?

    • Which format are you referring to?

      This isn’t meant to predict what people pay at auction – it’s meant to estimate a player’s worth. So if your league is way off from this, you could look at Estimated Value – ‘Market Price’ to determine best bargains (e.g., in leagues where pitchers are devalued, you may use $180/$80 or just aim to draft some pitchers at a few dollars less than predicted value.

      Negative values reflect players who shouldn’t be drafted in the format. Sometimes this is driven by the player just being substandard – sometimes it’s because their estimated playing time is too low and, thus, couldn’t start every day for a team.

  12. Dave says:
    (link)

    At home with the flu, working on my own spreadsheet based on a modified Mayberry Method, when I realized I need some projections to really get the feel for how my ranks will play out. I’ve been waiting for Steamer to release their projections and now that they’re a part of the best fantasy baseball website around… I think it’s time for me to junk my Mayberry sheet and use the brilliance that is RazzSteamer.

    • Glad to help. I don’t know everything about the Mayberry Method (BaseballHQ is subscriber-only) but it’s interesting. I’d like to see how it stacked up against other $ estimates – if BaseballHQ has an official version for ESPN 12 Team 5×5, I’ll add it into the 2013 projections test.

  13. BigPoppa says:
    (link)

    Rudy,
    I am very intrigued by Razzball. I am curious, would your ratings/values be applicable in our league? Or is there a way for me to edit your ratings based on the format/categories we score on? My league is a Yahoo 10-team H2H league that scores on:
    HITTING – 8 cetegories: AVG, OBP, Hits, Runs, HR, RBI, SB, Ks
    PITCHING: 7 categories: Wins, Losses, Saves, Holds, ERA, WHIP, Ks

    Thanks for all your help!
    -Jay

    • Intriguing? Man, I was going for beguiling!

      I think the 10 Team PS may be an okay starting point but there’s definite adjustments you’d need to make that are hard to quantify.

      For Hitting, the additional 3 categories (OBP, Hits, and Ks) all tend to reward slap hitters and penalize HR hitters without a good eye. I’d perhaps discount the HR Point Shares and double-credit the AVG Point Shares (to factor in Hits). Use the Steamer projections to account for OBP and Ks in some way.

      For pitchers, the addition of Losses and Holds devalues SPs and increase Middle Reliever value.

      From a strategy standpoint, I could see grabbing 2 ace SPs on good teams (aka not F-Her) that can give you Wins/ERA/WHIP/Ks with minimal L and then stock up on relievers. Your league is shallow enough that you can stream SPs if need be.

  14. Simply Fred

    simply fred says:
    (link)

    Rudy, why do I WASTE MY TIME? I have been grinding numbers for three months. I find that most of my targeted players match up very closely. AND, the few that I have been struggling with, you have slotted comfortably for me, so much so to–JUST RIDE RUDY! I will sleep much better tonight.

    (‘Bout time those closers got some respect!)

    • I imagine housewives felt the same way about unnecessary grinding upon the advent of the neighborhood butcher.

      Sleep well, Simply Fred.

  15. Deano says:
    (link)

    Pick 4 for 12-team mixed one catcher keeper

    Cargo $25
    Dickey $4
    Mauer $9
    Aramis Ram $9
    Butler $13
    Scherzer $9
    Desmond $4
    A. Hill $4
    Choo $12

    • A-Ram, Mauer, Butler, Desmond

  16. Simply Fred

    simply fred says:
    (link)

    What are the R/HR/RBI/SB/AVG for Nakajima? (they don’t show in the table)

    • @simply fred: I don’t have any yet. Steamer is working on the Asian imports (also includes pitchers Kyuji Fujikawa on the Cubs and Hyun-Jin Ryu on the Dodgers).

    • @simply fred: We (Steamer) haven’t translated Japanese stats yet but we will so Nakajima should be added in a later version. Offhand, I’d expect something in the neighborhood of .260/.330/.400 with ~15 HR and ~10 SB depending on playing time.

      • Simply Fred

        simply fred says:
        (link)

        @J. Cross: Thank you!

  17. Simply Fred

    simply fred says:
    (link)

    Rudy, have you targeted a 12-team draft? If so, what is the average $(160/100) (without bench)?

    • @simply fred: Grey is posting them soon. We’re leveraging the RCL league 2012 results for 12 team leagues. Since there aren’t major expected changes in run environment, I think those are better to use than what comes out of my model.

  18. Simply Fred

    simply fred says:
    (link)

    The next question is: what are the averages for hitting and pitching?

  19. Git Michaels says:
    (link)

    hey guys,
    i’m in a holds league for the first time this year. any chance these projections could include some holds in future versions?

  20. Simply Fred

    simply fred says:
    (link)

    Rudy, I don’t expect Moss to get more than the 460 PA, but can you tell us what his $(160/100) would be given 600 PA? Thank you!

    • @simply fred: He’d go from $0 to $10 although it’s probably less than that as he’s generally benched against Lefties so he’d fare worse in those extra 140 PA.

      • Simply Fred

        simply fred says:
        (link)

        @Rudy Gamble: As you note to AJA, Moss should come at a discount due to reduced playing time (bench against lefties). And, since I will be streaming and have more bench available, am considering running Moss. Not sure the $10 is enough to warrant it, but … He is so strong when he is in against righties I was thinking he would rate quite a bit higher than the $10 if extended to 600 PA…

  21. AJA says:
    (link)

    Hey Rudy, nice work on the player projections. By any chance has there been any work on quantifying and reporting the uncertainty or confidence intervals around these? In shallower leagues with deeper FA pools I find that if every player you draft meets their projection…you lose the league. You need guys to outperform and guys you can dump if they underpeform. My prior is that younger players, streakier players and guys coming off of injury are the most likely candidates for beating their projections, but it would be great to confirm that theory with actual data.

    • If u download Steamer’s projections, there is a confidence level for players. Since the projections represent the 50% case, it would mean everyone is as likely to underperform as over perform.

      I think it’s easier to target guys with strong peripherals who are undervalued because of reduced playing time. Guys off injury or down years usually come at discounted rates.

      You could filter the 2012 player rater based on $ diff between actual and predicted to see if u see something that will help u predict for 2013. But if it was easy to predict, projections would be more accurate.

  22. The Talented Mr. Dope Man says:
    (link)

    This is AWESOME! I used Greg’s rankings last year in conjunction w / steamer. This will make my process much easier… One small suggestion, would love if they would rank some minor leaguers that will hit the bigs this years.

    • Glad to help. We will have some post this preseason ranking rookies. So tough to predict playing time – let alone performance…

  23. Simply Fred

    simply fred says:
    (link)

    Pretty large gap between player-rater for Choo, #139 ranked, and where Grey has him, #14 OF. Rater pre-Cincy?

    • Nope. His projections factor in CIN. His 2012 stats look a lot worse if you factor in that he’ll likely play 10-15 games less since the reds have better 4th OF options.

      • Simply Fred

        simply fred says:
        (link)

        @Rudy Gamble: Thanks. You da’ MAN!

        • @simply fred: Thx. BTW, look at Choo’s platoon splits. He’s AWFUL vs. LHPs. That’s why I expect him to sit in favor of the squeezepack (Heisey) against LHPs (enough to lose 10-15 games). That would help his rate stats but only hurts counting stats.

          • Simply Fred

            simply fred says:
            (link)

            @Rudy Gamble: x’d choo off the target list; likely moss as well–due to pt issues. no sense in handicapping myself out of the shoot. Am sure I will need all available bench slots to compete.

  24. AJ says:
    (link)

    Rudy – I’ve got a question that seems like it’d be up your alley. Holding everything else constant, are DH-only players like Ortiz worth more in leagues with two utility spots than leagues with one? It would seem so instinctively but I’m curious how significant of a difference it makes in dollar values. Obviously it’s a different question than the difference in the value of catchers in one/two catcher leagues because you can fill the extra utility spots with non DH-only players.

    • @AJ: Everything equal, a DH’s lack of position eligibility for a DH has a lesser penalty on a 2 UTIL team (so, yes, more valuable). There is a value to having position-eligible guys in these roster spots (as you can slot them into other positions in daily leagues when your ‘starters’ are off and then fill their UTIL spot w/ anyone) but I’d be hard-pressed to even value that at $1.

      Generally, I’ve found that 2 UTIL leagues either axe the CI/MI, are 3-4 OF, or both. In those cases, DH value is lost solely because there are more quality hitters available for rostering then a typical CI/MI/5 OF league.

      FWIW, I’ve never come up with a satisfactory way to ‘penalize’ DHs for lack of position eligibility. For $ values and Point Shares, I basically just use 1B averages and multiply them up by 5%.

  25. Hi Rudy,

    How would you change this if you have 21 players? Our league uses the ESPN rosters except we have 1 If instead of 1 CI and 1 MI. How does this change things?

    • @jbird669: I think the biggest adjustment would be to 2B/SS as you’d be going from about 18 rostered players of each (12 2B, 12 SS, 12 MI) to something like 13. 1B and 3B would be affected less as any quality 1B and 3B bats would play at UTIL (knocking off some OFs).

      So I would probably take a few dollars off of 2B/SS and maybe a $1 off 3B.

      Another possible repercussion would be less league SBs since MI usually nets an above-average # of SBs. This would slightly reduce the value of an SB which would reduce the value of an SB guy and, conversely, increase the value of a slow guy at an SB-friendly position like OF.

  26. TheTinDoor says:
    (link)

    What is the “Reliability” Column an index of? Are they calculating somehow confidence in the projection?

    • @TheTinDoor: A player with no past performance data would have a reliability of 0 and a player with an infinite number of past plate appearances or innings or whatever would have a reliability of 1. One way to think about it is in terms of how much regression to the mean there is. A reliability of 0.80 means there past stats are regressed roughly 20%. This is an idea we took from Marcel’s reliability scores.

  27. TheTinDoor says:
    (link)

    Thanks for explaining the concept. Any more info you can give as to how roughly how it’s calculated? Based on your brief explanation above, it’s tied to AB/IP. So then you would expect players who are advanced in their career to have high reliability – and indeed, Miggy Cabrera, Derek Jeter, etc are at the top of the heap.
    But then you have 21-yr old Mike Trout at 0.83, just .04 below the top. Is your system looking at some type of historical comp – i.e., 21 year olds who have performed close to Trout generally produce at x-rate?
    On the other end of the spectrum, an established, older player like Chad Tracy comes in at a 0.59 reliability.

    Not trying to nitpick, just looking to get an understanding of how I can use that scale. Thanks,

    • @TheTinDoor: Reliability has a short-term memory so last year’s PA count decidedly more than the year before and so on. Trout got a ton of PA last year so his reliability is pretty high. That said, there is some distance between a reliability of .83 and one of .87. It might be more instructive to think about the amount of “unreliability” (1-reliability). Trout has something like 30% more unreliability Jeter, Cabrera et al.

  28. Simply Fred

    simply fred says:
    (link)

    I don’t see a number for AVG for Granderson. Is the overall correct? Or, does that need to be filled in (presumably with a negative number) which would downgrade his rating?

      • Simply Fred

        simply fred says:
        (link)

        @Rudy Gamble: So, his overall $21.40 didn’t change. The -.7 must have been factored in–just didn’t show…?

        • @simply fred: Yup, was already factored and represented in his Point Share total.

  29. Club Sauce says:
    (link)

    Rudy – I’m new to Point Shares but am very impressed (I’m sure you were working to impress me). I’d guess this gets asked a lot, but can I use Point Shares to evaluate a trade in a Keeper League?

    I’m considering offerring Rizzo (who can be kept in the 21st Round) and a 7th Round Pick for Heyward (who can be kept in the 14th Round) and a 16th Round Pick. Rizzo is projected at 1.2 Point Shares, and Heyward is projected at 2.0. The hitters ranked around 75 with our 36 Keepers removed on the Pt Shares list are projected at around -0.4. A 16th Round guy is projected at around -2.5 Pt Shares.

    Would I simply add all of the projected Pt Shares for the pieces involved? Do I also include the projected Pt Shares for the spot where each player would be kept? Doing this would seemingly unfairly harm a side of a trade receiving a late round pick.

    If I just look at a few of the pieces, the 0.8 difference in Heyward and Rizzo is larger than the Pt Shares for a player picked with the pick I’m losing (7th Round Pick: ~ -0.4). It just doesn’t seem accurate to think it’s that simple and seems wrong to ignore other variables involved in the trade (Keeper Value, 16th Round Pick I’d receive, etc.).

    Do we need to analyze Pt Shares differently when Keeper Values need to be taken into account?

    • I think Point Shares is a good way to evaluate any trade. It’s worth, noting, though, that you may want to discount Pitcher Points Shares about 5% in the Pre-Season since their value is less predictable than hitters (that’s already accounted for in the $).

      For keepers, the only additional aspect is the value of the player in subsequent years.

      To determine if the trade works for you, I’d calculate it as:

      Heyward + 16th round pick + Heyward’s future keeper value > Rizzo + 7th round pick + Rizzo’s future keeper value

      You do need to consider negative Point Shares as that 16th round pick will play for your team. Negative doesn’t mean a guy is unstartable – just that he’s below average.

      So 2.0 – 2.5 + Heyward’s future keeper value vs. 1.2 + -0.4 + Rizzo’s future value nets out to -0.5 for you and 0.8 for the other side. I think Heyward is a better bet for the future but 1.3 Point Shares is a pretty big gap. I’d look to either get a better pick in return or give up a weaker pick.

      • Club Sauce says:
        (link)

        @Rudy Gamble: Thanks, that’s very helpful.

        A few follow up questions, and feel to cut me off whenever necessary, as this all may be too specific to my situation to not be very relevant to other readers.

        Is your statement of 1.3 Pt Shares being a “pretty big gap” based on the general “unpredictability” of the future or a combination of both that and the two players’ keeper value in my specific case? I ask because 1.3 doesn’t necessarily seem like a ton, all things (keeper value, especially) being equal, especially considering Heyward has a predicted 0.8 advantage this season. If they both reach those projections, that only leaves him needing to gain 0.5 more Pt Shares than Rizzo over the course of his life as a keeper to give me the advantage, all things being equal, right? But, when “keeper value” is factored in, you may be saying it may be more difficult for Heyward to see that 1.3 Pt Share advantage since Rizzo is kept 7 rounds later and, thus, is deriving greater benefit for my team compared to Heyward in terms of the difference in each player’s projected Pt Shares and others drafted where the player is to be kept. If, and these are arbitrary values for the draft pick, but if a typical 14th rounder is projected for -2.5 Pt Shares, and a 21st rounder is typically projected for -4 Pt Shares, then my team may be drawing more benefit with Rizzo in terms of total Pt Shares “gained” each year even if Heyward is “outearning” RIzzo straight up by any margin less than 1.5 Pt Shares, right?

        Also, could you shine any light on how “Average” is defined for these values? Specifically, do these Pt Shares already have position completely factored in and, thus, not necessarily reflect “above average” for that specific position but “above average” for all players? For example, there are only 4 Catchers projected to be “above average” for 2013. Have those projected values already taken the Catcher position into account and, therefore, reflect that there are only projected to be 4 Catchers that are “above average” fantasy producers regardless of position? Is this where it is determined Catcher is the most valuable position, since there are typically less players projected to be “above average” fantasy producers each year? I ask because this comment: “Point shares represent a player’s estimated impact on an average team’s points versus the average drafted player at his position.” has me thrown off a bit. That could hint to me that there should roughly be as many players “above average” as there are “below average” based on roster size (in a 12 team league, 6 catchers above, 6 below), unless there are just a ton of “average” players, but those would still be reflected by a bunch of guys near 0.

        Again, this is all great stuff, thanks.

        • @Club Sauce: My ‘pretty big gap’ remark is based on the fact that, based on our best projections today, the deal wouldn’t pay off by end of next year. There may be an argument that Heyward is more of a sure thing than Rizzo and you’re buying less uncertainty. And, yes, I’m also factoring in that there is a bigger opportunity cost for keeping Heyward (14th round) vs. Rizzo (21st round).

          Point Shares are 75% based on the average player at the position and 25% at the overall average hitter. I found that mix worked best in most mixed league formats. So for Catchers, you’ll typically find less than 1/2 above 0.0 point shares because they are generally the weakest position – largely driven by reducing playing time estimates. But the 5th and 6th catchers in 12 Team ESPN are projected at -0.1 and -0.2 so it’s not so far out of whack.

          The system doesn’t try to force a 0.0 on the average hitter at a position. If there are 4 elite players and 12 so-so players at SS, it’s possible for the 4 elite players to drive up the averages so high that only those 4 are above 0.0. But looking at SS in 12 team ESPN, it looks like exactly 9 are >0 (12 SS + 6 MI) while 8 2B are > 0 and two are at -0.1.

          Hope that helps…

          • Club Sauce says:
            (link)

            @Rudy Gamble: Very helpful, indeed. Thanks again.

  30. harkrider says:
    (link)

    Rudy,

    trying to figure out which 5 to keep in a 10 team roto keeper league.

    Pujols, Votto, Heyward, Castro, Lawrie, Bruce, Medlen

    I’m leaning toward Pujols, Votto, Heyward, Castro and Bruce.

    Any thoughts?

    • @harkrider: Definitely Votto, Heyward, Castro, Bruce. Definitely no on Medlen. Pujols vs. Lawrie is a tougher call than it looks. I think Pujols has two more years in him where he’s a top 5 1B. Hard to know where Lawrie’s upside is. I’d consider trading either Pujols or Lawrie for a draft pick (if you’re league allows). Otherwise, I’d probably stick w/ Pujols…

      • harkrider says:
        (link)

        @Rudy Gamble: What about trading Pujols for either McCutchen, Braun, or J Upton? I think I could get those trades done, especially since I don’t need a 1B and 2 of those teams would like a strong 1B

  31. Dave says:
    (link)

    The RotoChamp software now has the RazzSteamer projections uploaded which is awesome. What are your thoughts on the composite projections RotoChamp offers (this would combine RazzSteamer, CAIRO and RotoChamp)

    Thanks!

    • I don’t recommend the composite after last year’s analysis (see bottom of the link below). Averaging two high-performing sources could work (steamer, Cairo) but adding rotochamp would be worse than using steamer or Cairo (if you reweigh the data against a real playing time source). At least according to how rotochamp performed last year…

      http://razzball.com/fantasy-baseball-projections-review-2012/

Comments are closed.