NLDS Game 3: Nationals 4, Giants 1 — The Nats’ offense was in doubt, to put it mildly. But anyone who expected a pitching mismatch hasn’t watched the last few AL Octobers.

Seven shut-down frames by Doug Fister ran his totals for eight postseason starts to 10 runs in 50.2 IP, and a 7-1 team record (4-1 personal). And today’s win avenged his one loss, by 2-0 to Madison Bumgarner in the 2012 WS (with the lone run charged to Fister being let in from first base by his relief in the 7th).

• Fister’s 1.78 runs per 9 IP is the best of the 91 pitchers with eight or more postseason starts. For all those with 40+ IP, only Mariano Rivera had a lower RA/9, 0.83. Christy Mathewson and George Earnshaw had better ERAs than Fister, but unearned runs push their RA/9 higher. (For seven or more starts, Sandy Koufax leads all 1.64 RA/9.)
• Of the 80 who’ve started at least five team wins, just four have as good or better a team win pct. as Fister’s 7-1: Lefty Gomez (7-0 team, 6-0 personal); Jerry Koosman (6-0, 4-0); Herb Pennock (5-0, 5-0); and Juan Guzman (7-1, 5-1).

Two mistakes on one play by Bumgarner broke up a scoreless duel in the 7th. Did the 0-and-2 bunt by Wilson Ramos catch the MadBum off guard? With no chance at the lead runner, he rushed a throw that skipped down the LF line and let in the game’s first two runs, and set up the third. It’s the kind of forced, desperate play you’d expect from a team facing elimination, not one in command of the series.

It feels hard to concede anything late in a scoreless playoff game, but the situation still called for caution — by the abstract Run Expectancy, and even more by the tenor of the series so far. Taking the free out would have left Nats on third and second with one down. But they hadn’t scored in their last 21 innings; they’d gone 1 for 16 with men in scoring position for the series; and the 8th spot and pitcher were due up. You have to make that team earn a run. If the sac does lead to a tally, you still have three times at bat (versus their two), with Fister probably pulled for a pinch-hitter, and an uncertain closer on tap.

…

Let’s have some fun with the Baseball-Reference Event Finder. How has Bumgarner fared when fielding a sac-bunt attempt grounder? Using the Event Finder for the regular season, plus a manual check of his postseason starts, I counted:

• 24 bunt grounders fielded
• 16 sacrifices (including one fielder’s choice with no out recorded)
• 3 singles
• 4 errors (3 scored as sacrifices)
• 4 forceouts at second base
• zero outs at third base

Out of four sac attempts with men on first and second, Bumgarner got the out at first base three times, and made one error that scored a run and fueled a big inning — an ill-advised throw to third, with Panda retreating to cover, none out and the game tied in the 8th.

__________

ALDS Game 3: Orioles 2, Tigers 1 — The better team advanced, with another close but convincing win.

“What was Buck Showalter thinking?”, giving an intentional walk to Nick Castellanos as the potential winning run with one out in the 9th? It’s just a rhetorical question, since we know what Buck told the mound assembly when announcing the IBB: “The next guy’s going to hit into a double play, and we’re gonna go home.” But he wasn’t just projecting confidence in a gut feeling. If you’re going to predict a DP, it helps to have these season stats in your pocket:

Highest Ground-Ball Rate, 1988-2014 (as a percentage of balls in play, 60+ IP):

Highest GIDP Rate, 2014 (60+ IP):

Britton’s DP rate was about three times the league average, 10.5% of all chances.

Another fact on Buck’s side: Castellanos is slow, scoring from first on a double just once in six chances this year, and the O’s had three good arms in the outfield. Absent a home run, it would probably take two hits to score Castellanos. And no one that Brad Ausmus could send up had much chance to go yard or string two hits together.

__________

NLDS Game 2: Giants 2, Nationals 1 (18 inn.) — Matt Williams lifted Jordan Zimmermann one out away from a 1-0 shutout, leaving the tying run aboard. That run came home, Zimmermann got no decision, and the Nats wound up losing.

We’re not here to question that move. (Joe Panik’s monster foul homer on 2-and-0, followed by two bad misses for a walk that brought up Buster Posey, had to cause some fear for Zimmermann’s future in those who watched the at-bat.) We just want to ask: Have you ever seen that before? Unless you’re a Cardinals fan of a certain age, probably not. That scenario had played out just twice in the last 96 years, both with Redbirds in the regular season:

May 29, 1992: Joe Torre pulled rookie Donovan Osborne from his 10th career outing, after Tony Gwynn’s two-out double on his 93rd pitch brought up Gary Sheffield. Lee Smith’s first pitch yielded a tying single, Fred McGriff doubled home Sheff for the lead, and Randy Myers struck out the side, ending with Andres Galarraga.

August 5, 1987: Whitey Herzog pulled veteran Bob Forsch after a two-out single. Two hits off Todd Worrell pulled the Expos even, and they won on a walk-off blast in the 13th. The tying and winning hits both came from Galarraga.

Sidebar: Osborne did finally bag a shutout, four years later, tossing 128 pitches under the watchful eye of first-year Cards skipper Tony LaRussa. He threw 125+ six times that year, including two of his last three in September, then got torched in the playoffs, and was never effective again. We imply no connection there, but perhaps in LaRussa’s evolution: There were 13 starts of 125+ pitches by Cardinals that year, nine in ’97, and no more than five in a subsequent season.

(There are two other games that could match our scenario, but lack play-by-play, in 1917 and 1915. Given the era, it’s quite unlikely that either game met our conditions, but you never know. And by the way, kudos to those who knew that Czech native Hugo Bezdek managed the Pirates from 1917-19.)

__________

AL Wild-Card Game: Royals 9, Athletics 8 (12 inn.) — No one told the Royals the game has changed since 1985, so they just went out and played like they did in that all-turf extravaganza.

Seven different Royals stole a base. That’s never happened before in the postseason, and just once in the regular season during the live-ball era.

• Naturally, that was done by the kleptomaniac 1976 A’s, who set live-ball records of 341 thefts, six different guys swiping 30+ and eight with 20+ — and nearly stole a sixth straight division crown. Ironically, after taking the lead in 12th inning on the legs of their 12th steal (also a searchable MLB record), the A’s lost when the first six Twins reached base in home half, with Stan Bahnsen walking Roy Smalley to force in the winning run.

Their seven total steals tied the postseason mark, set by Frank Chance’s Cubs in the 1907 Series opener — which wound up a tie, after Johnny Evers tried for a “steal-off” win in the 9th, and another potential game-winner was waved off for runner interference at home — and tied by the Big Red Machine in the ’75 NLCS.

• As noteworthy as the number of steals was the timing: The first four swipes came with the Royals trailing by at least two runs, including two in the 8th while behind by three or more. If Lorenzo Cain had been caught in that inning, there would have been two out and none on in a 7-4 game; with the 1st-inning misadventure, the game story would have been, “Royals Run Themselves Out of Postseason.” For what it’s worth, Tango’s Run Expectancy Matrix (1993-2010) values Cain’s steal at about one-sixth of a run, while the cost of a CS would have been almost half a run. Call them confident, cocky or cuckoo, they played their game to the end, and it worked.
• Guess which team had the only 7-SB game in this regular season.
• The last time Oakland allowed 7 steals was a special night for a catcher who allowed 155 steals the year before: The only 2-SB game of his career.
• And here’s the last time a team swiped eight in a game … and now you know why Victor Martinez became a DH.

Five thieves scored after stealing. Just one other postseason team had five players with both a steal and a run in the same game, and only three of those runs followed a steal.

• By the way, KC’s last steal of home was last September, by Alcides Escobar.

The Royals laid down four sacrifice bunts. One other team did that in a postseason game played with a DH. Alcides Escobar is the second ever with two hits and two sacs in a postseason game.

• Has the DH position ever been so engaged in postseason small-ball? Christian Colon pinch-sacrificed in the 10th for DH Terrance Gore … who had pinch-stolen in the 7th for DH Billy Butler … who had done this in the 1st.
• Colon is the first DH with a steal and a sac bunt in a postseason game. It’s also the first time two DHs stole a base for the same team.
• The flip side of the DH coin: Four DHs before Brandon Moss had hit 2 HRs in a postseason game, but only one of them had more than 2 RBI — in fact, that guy set a postseason record, since tied by three BoSox.

_____

Was Eric Hosmer involved in this game? Caught stealing home in the 1st … Caught Sam Fuld straying off second to end the 3rd … Placed a two-out flare for the lead in the 3rd … Walked in the 8th, ending Lester’s night (after 1-and-2, and a just-foul(?) bid for a hit-and-run double), and dashed home when Derek Norris lost sight of a wild pitch, cutting the gap to one run … Opened the 10th with an infield hit (0:30 here, and a bad night by Jed Lowrie on both sides), but died at third when Salvador Perez fell to 0-for-5 … Finally, Hosmer tripled to left-center, and charged home on a chop off the plate to tie it up in the 12th.

• Only 16% of Hosmer’s career batted balls have gone to the opposite field — but he’s hit .394 with a 1.019 OPS on those balls, with 45% of the hits earning extra bases, vs. 28% of his other hits. His ratio of OPS in that split vs. overall is second-highest of 99 active LHBs with at least 250 balls hit the other way.

_____

Many team characteristics played out as expected: The Royals ran, the A’s mashed. Oakland drew 6 walks to KC’s 3, each befitting its MLB rank (first and last). The Royals struck out just 8 times in 44 PAs, close to their MLB-best season rate. (But did you know that the A’s were 2nd-best in contact rate?) Thirteen singles exemplified KC’s MLB-high singles percentage (17.3% of PAs); they were the only team with more singles than strikeouts this year (+61), while the MLB average was minus-301. The A’s lost a one-run game; they were 21-28 in the regular year, worst of all the playoff teams.

• One key trait was betrayed: Oakland was the AL’s stingiest at allowing runs in the 7th-9th innings, but they surrendered four in those frames.

_____

Luke Gregerson logged his best ERA this year, but had his worst time with inherited runners, letting in 10 of 23. He also tossed a career-high six wild pitches; in those six games, he was charged with 7 runs in 6.2 IP, and let in all four inherited runners.

_____

In the last of the 12th, two out and a man on first, Bob Melvin pulled lefty Fernando Abad (who retired one man with one pitch) and brought in Jason Hammel to face the right-handed Perez. Did Melvin gain a platoon edge? To the extent we can judge this from stats, it comes down to how you weight yearly splits versus career — especially with Abad, who plucked this year’s 1.57 ERA out of a clear blue sky:

• Abad vs. RHBs — This year, .162 BA/.475 OPS … Career, .260 BA/.753 OPS
• Perez vs. LHPs — This year, .226 BA/.632 OPS … Career, .307 BA/.852 OPS
• Hammel vs. RHBs — This year, .233 BA/.670 OPS … Career, .271 BA/.740 OPS
• Perez vs. RHPs — This year, .270 BA/.710 OPS … Career, .278 BA/.712 OPS

Even if the move did gain a small platoon edge, was that enough to offset other factors?

• Abad squeezes the running game — 5 steals in 12 career tries. Hammel is runnable, averaging 19 SB and 6 CS per 200 IP (79%). With two out, it was a pretty good bet that Christian Colon would try to steal on Hammel.
• All other things being equal, Abad has a better chance to get one out: Both this year and career, his rates are better than Hammel’s across the board.
• Abad has been far better than Hammel with men on (career BA .262 vs. .299, OPS .722-.817), and especially with a man on first only (.218-.304 and .596-.800). With a man on second only — the situation after the steal — Hammel’s line is .308/.873.

It’s hard to fault Hammel’s work on Perez. He made a good pitch on 2-and-2, although a bit predictable, given how they’d attacked him all night with two strikes. But it wasn’t only the hit that beat Hammel. Colon’s steal played a big role, and the move to Hammel made that steal much more likely.

_____

My phone rang right after Brandon Moss crushed Yordano Ventura’s 98-MPH heater. My brother, Big Al, came right to the point: “Why? Why do managers think this will work in the postseason?”

“This” meant thrusting full-time starting pitchers into unaccustomed relief roles.

Now, we all remember some successes and failures in this department. But I wanted details. They’re not easy to harvest, so I had to draw tighter boundaries than I’d have liked. First, I gathered all pitchers with one or two relief outings* in any postseason from 1995-2012. (The 2013 results were then absent from the Play Index.) From that group, I found those who had made at least 20 starts that same year, and gathered those regular-season stats.

The overview of those 133 pitcher-seasons and their postseason relief work:

• Regular season averages: 4.13 ERA … 109 ERA+ … 29 starts … 183 IP
• Postseason relief totals: 4.74 ERA … 1.36 WHIP … 277 IP (1.65 IP/G) … 12 wins, 18 losses

But some of those guys were 5th starters who became postseason fifth wheels, maybe doing a bit of mop-up. We want the guys who were thought of as weapons. So I made a subset, the 70 pitchers with at least 25 starts and a 100 ERA+ in that regular season. Their numbers:

• Regular season averages: 3.64 ERA … 125 ERA+ … 31 starts … 202 IP
• Postseason relief totals: 4.03 ERA … 1.30 WHIP … 150 IP (1.78 IP/G) … 6 wins, 12 losses

That gap’s certainly not big enough to say it’s always a bad idea, but it’s more than enough to make you question the move to Ventura.

Some other high-profile failures:

• Kevin Brown, 1998 NLCS Game 5: With 3 days’ rest from a 122-pitch shutout in Game 2, Brown entered the 7th with a 2-run lead, one on and none out, Padres looking to close out the series. He got through that inning, but served a 3-run homer to Michael Tucker in the 8th, and Atlanta forced a Game 6. Brown had relieved twice out of 317 prior MLB games, yielding 3 runs in 3 IP. (Brown’s 41 pitches in Game 5 kept MLB’s overall WAR leader on ice, but Sterling Hitchcock matched zeroes with Tom Glavine for five innings, and the Pads put up five in the 6th to advance.)
• John Smoltz, 1999 NLCS Game 6: One of the best in postseason history, and his very first big-league relief game (after 380 starts) was a smooth save in Game 2, a one-run win. His next try came with 2 days’ rest from 7.1 innings worked in Game 4, and he started the 7th with a 7-3 lead as the Bravos looked to clinch. But the Mets crushed him with 4 hits in 5 batters, tying on Mike Piazza’s homer.
• Roy Oswalt, 2010 NLCS Game 4: With 2 days’ rest from a 111-pitch win in Game 2, Oswalt started the home 9th in a tie game against the heart of the order. He gave up the winning run on two hits and a sac fly, as the Giants seized a 3-1 series lead. Charlie Manuel used four relievers before Oswalt, but not his $12 million closer, Brad Lidge. Oswalt had relieved three times in the last six years, yielding 2 runs on 4 hits out of 10 batters.
• Jack McDowell, 1995 ALDS Game 5: Black Jack’s first relief work after 224 big-league starts came in a tight spot in a do-or-die game: Tied in the home 9th, two on and one out. He got past Edgar and A-Rod and dodged his own bullet in the 10th, but blew a one-run lead in the 11th on Edgar’s 2-run double with none out, making 5 hits out of 10 batters. McDowell was rocked in Game 3 (and both starts in the ’93 ALCS), and had one day’s rest from that 85-pitch labor; his postseason record is a stain on a fine career.
• Charles Nagy, 1997 WS Game 7: Second relief outing out of 200+ games, and the first since his 1990 rookie season.

Perhaps someone out there would like to dig up the high-profile successes?

* Why did I take just those with one or two relief outings in a postseason? There are more than 1,000 October “reliever-seasons” from 1995-2012; most are just noise to this inquiry, and wading through the whole lot would have been too much work. We’re after the one-offs — the Big Units (three relief games, over three different seasons), not the Marianos. I made a guess that a ceiling of two relief games per postseason would give us most of the relevant data.