Free Agent Class of 2019

The official MLB site, released a list of the free agent class of 2019.  Let’s run that list through this data model.

There are many ways to sort this list and calculate career numbers.  During the season this model ranks players according to total WAA value for that season.  Number of years can differ among free agents.  It was decided to simply use WAA valuations for the last 3 seasons.   This is a pretty good indicator showing what have you done for me lately.

The below table shows the top 25 free agents sorted by total WAA for their last three year splits.  The first column shows rank for the 2018 season which just concluded.

Top 2019 Free Agents

2018 Rank 3 Year WAA Name_TeamID Pos
+027+ 14.68 Bryce_Harper_WAS RF-CF
+076+ 14.24 Nelson_Cruz_SEA DH
XXXXX 11.51 Daniel_Murphy_TOT 2B-1B
XXXXX 9.39 Brian_Dozier_TOT 2B
XXXXX 9.37 Josh_Donaldson_TOT 3B-DH
+158+ 9.03 J.A._Happ_TOT PITCH
XXXXX 8.97 Andrew_Miller_CLE PITCH
+083+ 8.27 Manny_Machado_TOT SS-3B
XXXXX 8.02 Zach_Britton_TOT PITCH
XXXXX 7.67 Adrian_Beltre_TEX 3B-DH
+163+ 7.48 Craig_Kimbrel_BOS PITCH
+148+ 6.91 Carlos_Gonzalez_COL RF
XXXXX 6.87 Mark_Reynolds_WAS 1B-3B
XXXXX 6.72 Brad_Brach_TOT PITCH
+135+ 6.72 Matt_Adams_TOT 1B-LF
-093- 6.57 Ervin_Santana_MIN PITCH
XXXXX 6.49 David_Robertson_NYA PITCH
+051+ 6.38 Charlie_Morton_HOU PITCH
+119+ 6.24 Evan_Gattis_HOU DH
+181+ 5.02 Dallas_Keuchel_HOU PITCH
+049+ 4.93 Hyun-Jin_Ryu_LAN PITCH
+199+ 4.60 Jeurys_Familia_TOT PITCH
+154+ 4.47 DJ_LeMahieu_COL 2B
XXXXX 4.45 Gio_Gonzalez_TOT PITCH
+194+ 4.39 Kelvin_Herrera_TOT PITCH

There is a lot more that goes into the calculus of deciding upon a free agent than simple run production.  Daniel Murphy who the Cubs acquired is ranked 3rd above but had a mediocre season according to his 2018 rank.  Does he have anything left in the tank going forward?   Ditto for Josh Donaldson,  Brian Dozier, and Andrew Miller.  This data model cannot predict the future.  The above simply shows a factual representation of how many wins each player contributed to their teams the last three seasons.

And as always, past results don’t affect future results.  They only show capability.  That is all for now.   A Giancarlo Stanton article is forthcoming.   Also working on historical daily simulations to compare them with historical vegas lines and historical Nate Silver predictions.  Until then ….

2018 World Series Report Part 5

With Boston up 3-1 this could be the last World Series report until 2019.

DATE 10_28_8:15_PM BOS LAN

LINEAWAY BOS [ 0.465 ] < 0.435 > +130 $230
STARTAWAY 2.21(0.556) David_Price_BOS TIER 3
LINEHOME LAN [ 0.556 ] < 0.583 > -140 $171
STARTHOME 5.02(0.640) Clayton_Kershaw_LAN TIER 1

BOS Lineup 1 ==> LAN Starter 1 / Relief 2 == 0.481 BOS 4.45 runs
LAN Lineup 1 ==> BOS Starter 3 / Relief 3 == 0.519 LAN 4.65 runs

Tier Combo 111 89
Home Field 106 92

Kershaw last pitched against Chris Sale in Boston for Game 1 so this match up is different.  Lines about the same as yesterday but TC Simulation has Dodgers favored at 0.519 break even probability.  This lowers Expected Value for Boston to 111 on a 100 risk which is close to their basic Home Field disadvantage Expected Value.

This could be the last game of the season.  If not, part 6 will be forthcoming.

Update:  It looks like Nate Silver’s model agrees.  Here’s a snapshot in case that link gets broken.


Update2: I’m an idiot again! It might be dyslexia but read the above table opposite of what it says. Nate had BOS at 52%.  TC Simulation had Dodgers at 52%.  Nate’s model would have generated an Expected Value for Boston at 120 on a 100 bet making it a betting opportunity.  Since we’re from the future it would have been a successful betting opportunity.  Home Field Expected Value would still be 106 however and how that factors into all of this is still a work in progress.

More historical analysis between vegas lines, TC simulation, and The above coming soon in the off season ( like now ).  This will start with 2011 – 2017.  Current year event files are needed for this analysis and usually releases them mid December some time.   </ end of update>

2018 World Series Report Part 4

Another World Series Report after not being able to watch the end of Game 3 because the bar closed.

DATE 10_27_8:05_PM BOS LAN

LINEAWAY BOS [ 0.400 ] < 0.426 > +135 $235
STARTAWAY 0.92(0.532) Eduardo_Rodriguez_BOS TIER 3
LINEHOME LAN [ 0.615 ] < 0.592 > -145 $168
STARTHOME 1.26(0.543) Rich_Hill_LAN TIER 3

BOS Lineup 1 ==> LAN Starter 3 / Relief 2 == 0.507 BOS 4.86 runs
LAN Lineup 1 ==> BOS Starter 3 / Relief 3 == 0.493 LAN 4.79 runs

Tier Combo 119 83
Home Field 108 91

Tier Combo simulations favor Boston very slightly due to  starting pitching evening out.  Lines are about the same as yesterday.  Simulation has no way of knowing how badly hurt each team’s relief staff is due to the marathon game yesterday.

2018 World Series Report Part 3

Game 3 starts tonight in LA so it’s time for another World Series Report.

DATE 10_26_8:05_PM BOS LAN

LINEAWAY BOS [ 0.408 ] < 0.408 > +145 $245
STARTAWAY -0.73(0.483) Rick_Porcello_BOS TIER 3
LINEHOME LAN [ 0.608 ] < 0.608 > -155 $164
STARTHOME 4.64(0.652) Walker_Buehler_LAN TIER 1

BOS Lineup 1 ==> LAN Starter 1 / Relief 2 == 0.474 BOS 4.61 runs
LAN Lineup 1 ==> BOS Starter 3 / Relief 3 == 0.526 LAN 4.89 runs

Tier Combo 116 86
Home Field 113 89

Last season these simulations were a concept and a pattern kind of emerged during those playoffs.  In the off season we’ll be able to do these daily reports for past playoffs using past data to see if similar patterns exist.

Simulations build upon summation of talent in various categories and since tiering uses a distribution from all 30 MLB teams, top teams are going to have top tier talent making it hard to discern a difference mathematically.

Tier Combo simulation above favors the Dodgers with the biggest edge of the series so far at 0.526.  This is largely due to differences in starters, Tier 1 vs. Tier 3, and difference in relief, Tier 2 vs. Tier 3.

Lines for today  favor the home team at 0.608 giving Boston a rather high Expected Value of 116 on a 100 bet.  Home field disadvantage drops them to 113 which isn’t much.  Because of razor thin difference in talent, either home field advantage or home field fan irrational exuberance is pushing EV up for visiting teams during these playoffs.  We observed LAN home field irrational exuberance when they played the Cubs.  Since World Series are national, LAN bettors should have less of an impact on their line.

Simulations do not use hard boundaries and from  lineup tables in World Series Report Part 2, Boston has around twice the lineup value as Dodgers ( 29 – 16 ), yet the sum of hitting talent is almost the same.

TeamID W-L Total Hitters Pitchers Starters Relief
LAN 21 50.7 28.1 22.6 14.9 7.6
BOS 54 42.2 31.6 10.6 8.9 1.7

Let’s see how those two numbers are calculated.  Below are batting rosters for each team.

Hitting Rosters


Rank WAA Name_TeamID Pos
+002+ 10.08 J.D._Martinez_BOS DH-LF-RF
+010+ 7.54 Mookie_Betts_BOS RF-CF
+035+ 4.77 Xander_Bogaerts_BOS SS
+039+ 4.45 Andrew_Benintendi_BOS LF-CF
+131+ 2.35 Mitch_Moreland_BOS 1B
+152+ 2.08 Steve_Pearce_TOT 1B-DH-LF
XXXXX 1.64 Rafael_Devers_BOS 3B
XXXXX 1.62 Jackie_Bradley_Jr._BOS CF-RF
XXXXX 0.53 Brock_Holt_BOS 2B-SS-RF
XXXXX -0.02 Blake_Swihart_BOS CR-DH-RF-1B
XXXXX -0.82 Ian_Kinsler_TOT 2B
XXXXX -1.28 Eduardo_Nunez_BOS 2B-3B-DH
XXXXX -1.30 Sandy_Leon_BOS CR
Total 31.64

J.D. Marinez should be AL MVP with Mookie Betts second.  There are 4 batters above in top 100 and 6 in the top 200.  Let’s look at the Dodgers.

Los Angeles

Rank WAA Name_TeamID Pos
+028+ 4.98 Max_Muncy_LAN 1B-3B-2B
+057+ 3.67 Matt_Kemp_LAN LF-RF
+083+ 3.07 Manny_Machado_TOT SS-3B
+109+ 2.58 Yasiel_Puig_LAN RF
+127+ 2.42 Joc_Pederson_LAN LF-CF
+149+ 2.10 Cody_Bellinger_LAN 1B-CF
+155+ 2.06 Justin_Turner_LAN 3B
+174+ 1.91 Yasmani_Grandal_LAN CR
+179+ 1.87 David_Freese_TOT 3B-1B
+196+ 1.74 Enrique_Hernandez_LAN CF-2B-SS-RF-LF
XXXXX 1.49 Chris_Taylor_LAN SS-CF-LF-2B
XXXXX 0.92 Brian_Dozier_TOT 2B
XXXXX -0.71 Austin_Barnes_LAN CR-2B
Total 28.1

Only two guys in the top 100 and 9 batters in top 200.  Boston has top of the league hitters, Dodgers have a lot of upper middle of league hitters.  Boston can put more value in a single lineup because talent is so concentrated and so high.

Simulations rely on differences with a cap.  This will be explained more in off season.  Right now Boston fields a Tier 0 lineup which is usually a tier or two above most high value pitchers they face.   This depresses LAN break even probability from TC simulation and increases BOS.

Today, even with such a massive lineup Boston brings, both TC simulation and lines favor Dodgers.  Dodger bettors are paying a large premium to bet the LAN line.  Right now Dodger bettors think their team has a greater than 0.608 probability of winning.    Nate Silver has LAN at 0.560 probability.

That is all for now.  A shorter Game 4 report will come tomorrow.  Until then ….

2018 World Series Report Part 2

Another day, another World Series Report.

DATE 10_24_8:05_PM LAN BOS

LINEAWAY LAN [ 0.444 ] < 0.435 > +130 $230
STARTAWAY 4.03(0.720) Hyun-Jin_Ryu_LAN TIER 1
LINEHOME BOS [ 0.574 ] < 0.583 > -140 $171
STARTHOME 2.21(0.556) David_Price_BOS TIER 3

LAN Lineup 1 ==> BOS Starter 3 / Relief 3 == 0.503 LAN 4.65 runs
BOS Lineup 1 ==> LAN Starter 1 / Relief 2 == 0.497 BOS 4.61 runs

Tier Combo 116 85
Home Field 106 92

Yesterday LAN was around 118 Expected Value on a 100 bet and today it’s almost the same.   As we saw in the playoff horse race series, playoff teams have the best collective talent and usually, including this season, the top top tier talent end up in the World Series.

TC Simulations are based upon regular season results where  elite teams face the dregs of MLB.  Top teams facing top teams usually end up being a 50/50 event historically.   Today LAN eeks out BOS at 0.503 but for all intents and purposes that’s a flip of a coin.  This is where home field advantage comes into play.  Boston is slightly over valued based upon that.  Since it’s a cold day that should favor Boston as well.

If lineups become available they’ll post below in an update.

Update mid game2:  Lineups are finally available.  Our lineup source was down for awhile so  estimated lineups based upon snapshots were used.  BOS fielding a slightly better lineup than the snapshot.

LAN Lineup

WAA Name_TeamID Pos PA 10242018
1.03 Brian_Dozier_LAN 2B 632
2.06 Justin_Turner_LAN 3B 426
1.78 David_Freese_LAN 3B-1B 312
3.04 Manny_Machado_LAN SS-3B 709
1.49 Chris_Taylor_LAN SS-CF-LF-2B 604
3.67 Matt_Kemp_LAN LF-RF 506
1.74 Enrique_Hernandez_LAN CF-2B-SS-RF-LF 462
2.58 Yasiel_Puig_LAN RF 444
-0.71 Austin_Barnes_LAN CR-2B 238
Total WAA=16.70 PA=4333 WinPct=0.574

BOS Lineup

WAA Name_TeamID Pos PA 10242018
7.54 Mookie_Betts_BOS RF-CF 614
4.45 Andrew_Benintendi_BOS LF-CF 661
2.12 Steve_Pearce_BOS 1B-DH-LF 251
10.08 J.D._Martinez_BOS DH-LF-RF 649
4.77 Xander_Bogaerts_BOS SS 580
1.64 Rafael_Devers_BOS 3B 490
-0.86 Ian_Kinsler_BOS 2B 534
1.62 Jackie_Bradley_BOS CF-RF 535
-2.10 Christian_Vazquez_BOS CR 269
Total WAA=29.25 PA=4583 WinPct=0.623

That might be the highest valued set of 9 players for any team this season.