2019 Playoff Horse Race Part 10

This will be the last regular season playoff horse race post during regular season.

ELO

elo0916

The above is a screenshot taken today (9/16) of 538’s ELO model.  Don’t know how ELO is compiled or what a “Team Rating” number represents.  ELO has OAK and CLE now in top 5 with WAS bumped down to mid tier.  Since ELO does not break its numbers out like we do below there isn’t much to comment on.

WAA

TeamID W-L Total Hitters Starters Relief UR
LAN 43 63.3 25.9 28.3 9.1 -15.6
HOU 45 58.6 29.1 26.8 2.8 12.4
NYA 45 41.5 34.0 -0.7 8.1 12.4
MIN 33 41.2 26.5 4.9 9.8 -14.6
WAS 16 37.0 16.8 18.8 1.4 17.4
ATL 35 35.5 17.5 13.5 4.5 -10.6
OAK 30 31.2 18.0 7.2 6.0 22.4
CHN 13 27.0 10.4 9.8 6.8 0.4
CLE 24 24.4 1.3 15.8 7.3 3.4
SLN 17 22.1 -2.2 13.5 10.8 10.4
BOS 9 17.1 20.2 -4.2 1.1 0.4
NYN 5 13.3 1.3 18.4 -6.5 10.4
TBA 27 12.7 -4.9 10.7 7.0 2.4
PHI 4 5.3 5.4 0.2 -0.3 0.4
MIL 11 0.7 1.8 5.8 -6.9 8.4

The above lists top 15 teams according to real team WAA (W-L) and then sorted on Total WAA according to this data model.  The numbers above are based on current rosters.  Since we’re in expanded roster month there may be some distortions up or down that may not represent  an actual 25 player playoff roster.  This model only counts stuff from current rosters and cannot discern who is there because of expansion.  The WAR table below is counted the same way.

WAA and WAR below has Cubs mid third; ELO has them low third.  All three models now have HOU and LAN top two and all relegate ATL to middle of the pack.  Cubs are on a sugar high with hitting because of their last three games with PIT where they put up double digit runs each game.  This propelled them ahead of SLN even though Cardinals are ahead in the only stat that matters in baseball.   Pitching has boosted SLN into first place in NL Central.

Unearned Runs above average (UR) has runs as units.  All other columns are WAA values  with wins as units.   LAN still has the worst UR but it’s up from past parts to this series.

WAR

TeamID W-L Total Hitters Starters Relief UR
HOU 45 64.4 37.9 21.3 5.2 12.4
LAN 43 47.8 32.5 13.2 2.1 -15.6
OAK 30 47.4 30.6 11.2 5.6 22.4
NYA 45 46.9 33.9 6.4 6.6 12.4
WAS 16 46.2 21.6 20.1 4.5 17.4
MIN 33 42.9 29.9 7.3 5.7 -14.6
ATL 35 42.5 22.6 14.3 5.6 -10.6
CHN 13 42.1 22.8 12.7 6.6 0.4
BOS 9 41.6 24.2 7.9 9.5 0.4
TBA 27 38.2 22.3 10.1 5.8 2.4
CLE 24 37.1 17.6 13.3 6.2 3.4
SLN 17 36.9 21.4 10.3 5.2 10.4
NYN 5 35.2 18.1 17.4 -0.3 10.4
MIL 11 28.4 20.2 7.7 0.5 8.4
PHI 4 27.4 17.9 7.0 2.5 0.4

WAR has HOU way above 2nd place LAN in Total, so much so it could be its own tier.  These WAR and WAA tables are based on current rosters with values simply added together in each category.  HOU has best starting rotation in WAR, WAA has LAN with the best.  BOS has the best set of relievers according to WAR, WAA has SLN as the best.

Let’s spot check these numbers on where WAR and WAA differ most.  Both systems agree HOU has top notch pitching close enough to be tied for first.  WAR has LAN starting rotation mid tier.   In past parts to this series WAR has been shown to under value LAN pitching; both relief and starters.  Let’s look at starters.

LAN Starters WAA

Rank WAA Name_TeamID IP
+003+ 8.44 Hyun-jin_Ryu_LAN 168.7
+025+ 5.56 Clayton_Kershaw_LAN 165.3
+028+ 5.35 Walker_Buehler_LAN 171.3
+096+ 3.13 Julio_Urias_LAN 74.0
+144+ 2.27 Rich_Hill_LAN 53.7
+159+ 2.16 Ross_Stripling_LAN 84.7
XXXXX 1.36 Tony_Gonsolin_LAN 35.0
Total 28.27 0.669

Dodgers list 7 pitchers as starters which is a lot.   HOU follows closely in second for starter value and they only list 4 which is a typical playoff starting rotation.  Dodgers have 4 starters in top 100 which phenomenal.  Houston has three starters in the top ten which is even better according to this data model.

LAN Starters WAR

Rank WAR Name_TeamID
+045+ 4.5 Hyun-Jin_Ryu_LAN PITCH
+096+ 3.2 Clayton_Kershaw_LAN PITCH
+180+ 2.1 Walker_Buehler_LAN PITCH
XXXXX 1.0 Julio_Urias_LAN PITCH
XXXXX 0.9 Rich_Hill_LAN PITCH
XXXXX 0.8 Ross_Stripling_LAN PITCH
XXXXX 0.7 Tony_Gonsolin_LAN PITCH
Total 13.2

According to rank, WAR sees far less value in LAN starters than WAA.  How can that be when LAN still ranks far ahead of MLB in PITCH according to seasonal runs scored against.

Top 5 PITCH

Tm W L BAT PITCH UR
LAN 97 54 95.1 159.8 -15.6
CLE 87 63 -19.9 118.9 3.4
SLN 83 66 -32.9 112.9 10.4
TBA 89 62 -9.9 111.8 2.4
HOU 98 53 138.1 105.7 12.4

These are the top 5 MLB teams according to PITCH which is a seasonal value.  There is no way Dodgers starters can be rated so low according to WAR having such high seasonal  PITCH.   Although current rosters can deviate from seasonal numbers, that’s not the case here.  All 5 of these teams should have top tier pitching ( relief + starters ) and they do according to WAA.  WAR has LAN ranked in bottom third of playoff teams with respect to total pitching which is absurd.

WAR under values relief as well.   Both BOS and SLN have major differences with WAR.

SLN Relief WAA

Rank WAA Name_TeamID IP
+058+ 3.82 Giovanny_Gallegos_SLN 67.3
+138+ 2.31 John_Brebbia_SLN 68.7
+192+ 1.85 Ryan_Helsley_SLN 31.0
XXXXX 1.62 John_Gant_SLN 64.3
XXXXX 1.51 Carlos_Martinez_SLN 41.7
XXXXX 0.69 Daniel_Ponce_de_Leon_SLN 45.7
XXXXX 0.48 Andrew_Miller_SLN 47.7
XXXXX 0.40 Tyler_Webb_SLN 49.0
XXXXX 0.29 Junior_Fernandez_SLN 8.7
XXXXX -0.10 Genesis_Cabrera_SLN 16.7
XXXXX -0.90 Dominic_Leone_SLN 36.7
XXXXX -1.16 Mike_Mayers_SLN 16.7
Total 10.81 TIER=3.24

Tiers are calculated based upon the top 8 relievers.  Cardinals will have a very good relief staff going into the playoffs according to this data model. At Tier +3.24 SLN is over 1.5 standard deviation above league average for relief.

SLN Relief WAR

Rank WAR Name_TeamID
+180+ 2.1 Giovanny_Gallegos_SLN PITCH
XXXXX 1.0 John_Brebbia_SLN PITCH
XXXXX 0.8 John_Gant_SLN PITCH
XXXXX 0.8 Carlos_Martinez_SLN PITCH
XXXXX 0.7 Ryan_Helsley_SLN PITCH
XXXXX 0.5 Daniel_Ponce_de_Leon_SLN PITCH
XXXXX 0.3 Tyler_Webb_SLN PITCH
XXXXX 0.2 Junior_Fernandez_SLN PITCH
XXXXX -0.2 Andrew_Miller_SLN PITCH
XXXXX -0.2 Genesis_Cabrera_SLN PITCH
XXXXX -0.4 Dominic_Leone_SLN PITCH
XXXXX -0.4 Mike_Mayers_SLN PITCH
Total 5.2

WAR only has one SLN reliever barely ranked in top 200.  One would think this is merely a so so relief staff just looking at the above numbers.

BOS Relief WAA

Rank WAA Name_TeamID IP
+060+ 3.78 Brandon_Workman_BOS 66.3
+189+ 1.85 Marcus_Walden_BOS 74.0
XXXXX 1.60 Josh_Taylor_BOS 45.3
XXXXX 0.55 Darwinzon_Hernandez_BOS 29.3
XXXXX 0.50 Matt_Barnes_BOS 59.0
XXXXX 0.19 Andrew_Cashner_BOS 142.3
XXXXX -0.02 Colten_Brewer_BOS 50.0
XXXXX -0.15 Travis_Lakins_BOS 18.7
XXXXX -0.55 Josh_A._Smith_BOS 29.0
XXXXX -0.63 Bobby_Poyner_BOS 8.0
XXXXX -0.65 Ryan_Weber_BOS 36.0
XXXXX -1.11 Ryan_Brasier_BOS 49.7
XXXXX -1.11 Hector_Velazquez_BOS 51.7
XXXXX -1.13 Mike_Shawaryn_BOS 17.3
-126- -2.06 Jhoulys_Chacin_BOS 94.3
Total 1.06 TIER=0.99

According to this model BOS relief, at Tier 0.99, are about 1/2 standard deviation above league average but well below the Cardinals’ +3.24 tier number.  Let’s see what WAR says.

BOS Relief WAR

Rank WAR Name_TeamID
+119+ 2.8 Brandon_Workman_BOS PITCH
+126+ 2.7 Andrew_Cashner_TOT PITCH
XXXXX 1.4 Marcus_Walden_BOS PITCH
XXXXX 1.3 Josh_Taylor_BOS PITCH
XXXXX 0.9 Matt_Barnes_BOS PITCH
XXXXX 0.6 Colten_Brewer_BOS PITCH
XXXXX 0.3 Darwinzon_Hernandez_BOS PITCH
XXXXX 0.2 Ryan_Weber_BOS PITCH
XXXXX 0.1 Hector_Velazquez_BOS PITCH
XXXXX 0.1 Josh_Smith_BOS PITCH
XXXXX 0.1 Travis_Lakins_BOS PITCH
XXXXX -0.2 Jhoulys_Chacin_TOT PITCH
XXXXX -0.2 Bobby_Poyner_BOS PITCH
XXXXX -0.2 Mike_Shawaryn_BOS PITCH
XXXXX -0.3 Ryan_Brasier_BOS PITCH
Total 9.6

SLN relief had a total WAR=5.2 which is far less that the above 9.6.  WAR thinks BOS Relief >> SLN Relief.  This model has an opposite view where SLN >> BOS.  Which model is correct?

The next playoff horse race will be posted when the 8 teams in NLDS and ALDS release their 25 player rosters.  The two wild card games will be handicapped individually and there will be some historical playoff handicapping leading up to the actual playoffs.  Final Cubs status for regular season coming in a few days.  Until then ….