Numbers Don't Lie
Sabermetrics Deep Dive
I am not going to discuss how my predictions from the last column fared. Not today. We have more important work to do—specifically, the work I should have been doing since the first week of the season, the work that explains more about this pennant race than Pythagorean expectation, more than ballpark factors, more than the gap between ERA and RCERA, more than any of the analytical tools I've been leaning on for four columns.
Defense.
Not the abstract concept of defense. Not the narrative shorthand of "good glove" or "smooth hands." The actual, granular, measurable machinery by which a ground ball to the left side either becomes an out or a single, by which a line drive to the gap either dies in an outfielder's glove or rattles to the wall for a double, by which a runner on first either stays at first on a single to right field or motors into third because the right fielder's arm is a suggestion rather than a weapon.
Diamond Mind Baseball—the simulation engine that drives every game in the World League of Baseball—models all of this with a specificity that most owners in our league have never fully interrogated. I hadn't either, until I started pulling at the thread of Iron Knob's pitching dominance and realized that the thread led not to the mound but to the men standing behind the pitcher. What I found changed how I understand the pennant race—and it should change how every owner in the WLB builds a roster.
Diamond Mind rates every fielder on three separate axes, each of which operates independently in the simulation.
Range is the most important. It measures a player's ability to reach batted balls and convert them into outs—not just the ability to cover ground, but the full package of positioning, quickness, soft hands, quick release, arm strength, accuracy, and decision-making that Diamond Mind's designers call "playmaking ability." The scale runs from Excellent to Very Good to Average to Fair to Poor. This is not a cosmetic label. When a ground ball is hit to the shortstop side of second base, the simulation checks the shortstop's range rating to determine whether he gets to the ball. An Excellent-range shortstop will convert a measurably higher percentage of those batted balls into outs than a Poor-range shortstop. The balls he doesn't reach become singles. Some become doubles. The pitcher's ERA goes up, his opponent batting average climbs, his WHIP inflates—and none of it had anything to do with the pitch he threw.
Error rate is expressed as a percentage relative to the era-average at that position. A rating of 100 means the player commits errors at the average rate for his position. A rating of 50 means he boots it half as often. A rating of 200 means he's throwing balls into the stands at twice the expected rate. Errors are dramatic—they put runners on, advance baserunners, extend innings. But here's the critical insight from the research: error rate is substantially less important than range at every position on the diamond. A shortstop who gets to everything and boots a few is more valuable than a shortstop who only reaches the routine balls and fields them cleanly. The engine knows this. The engine has always known this.
Throwing applies to outfielders and catchers. For outfielders, the arm rating determines whether runners take extra bases on singles, doubles, and fly balls—whether the man on first goes to third on a single to right, whether the runner on second tags and scores on a fly ball to center. For catchers, throwing governs stolen base outcomes. These ratings are not abstract. They fire on every relevant play. A right fielder with a Poor arm is functionally giving the opposing offense extra bases all game long—bases that translate into runs, runs that translate into losses.
The simulation combines all of these factors—range, errors, throwing—with the pitcher's own ratings and the ballpark dimensions to produce the outcome of every batted ball. The pitcher and batter contribute equally to the plate appearance result. But once the ball is put in play, the fielder takes over. And this is where the pennant race lives.
The most rigorous study of Diamond Mind's fielding impact—a massive simulation project conducted by a community researcher, testing every combination of range and error ratings across fifty seasons per scenario—produced numbers that should make every owner in the WLB reconsider their roster construction.
At shortstop, the difference between an Excellent-range player with a solid error rate and a Poor-range player with the same error rate is approximately 25 to 30 runs over a full season. Twenty-five to thirty runs. In a league where the gap between Iron Knob's +79 run margin and Knockemstiff's +49 is thirty runs, the difference between an elite shortstop and a poor one is worth more than the gap between first place and fifth place.
The numbers get more revealing when you examine how range and errors interact. A shortstop rated Poor with an error rate of 70—meaning he doesn't get to many balls, but he fields the ones he reaches cleanly—costs his team roughly eight runs over a season compared to the Average/100 baseline. That's manageable. That's a roster construction tradeoff you can absorb if the bat compensates. But a shortstop rated Poor with a league-average error rate of 100 costs roughly seventeen runs. And a Poor shortstop who also boots them at a 125 rate? The damage escalates beyond twenty runs, approaching the territory where you're simply donating two wins to every opponent who puts the ball on the ground.
Here's the finding that should fundamentally alter how you think about fielding at short: an Excellent-range shortstop with a grotesque error rate of 200—a player who gets to everything and then throws it into the photographer's well twice as often as normal—is approximately equivalent defensively to a Poor-range shortstop with a clean 70 error rate. The range advantage is so large that it compensates for catastrophic hands. The engine is telling you something that Bill James argued decades ago and that major league front offices didn't fully internalize until the Statcast era: the plays you don't make matter more than the plays you bobble.
An Average-range shortstop with a 75 error rate—a solid, unremarkable defender—is roughly equivalent to an Excellent shortstop with a 125 error rate. The simulation doesn't care about your fielding percentage. It cares about how many batted balls you convert into outs, by whatever means necessary.
Now look at the WLB's shortstops through June 6 and tell me this doesn't matter.
Ozzie Smith is playing 55 games at shortstop for Iron Knob: 79 putouts, 228 assists, 8 errors, 41 double plays, .975 fielding percentage. That's 307 total chances handled. Cal Ripken is playing 57 games for Nicaragua: 70 putouts, 176 assists, 4 errors, 21 double plays, .984 fielding percentage. Ripken's fielding percentage is nine points higher than Smith's. He's also handled 81 fewer total chances. Eighty-one fewer plays in virtually the same number of games. Ripken has better hands. Smith gets to more balls. The simulation doesn't care about the fielding percentage gap. It cares about the 81-play gap. Every ball Smith reaches that Ripken doesn't is a ball that becomes a single against Nicaragua and an out against Iron Knob. Multiply that across 154 games and the run impact is enormous.
Greg Gagne has started all 60 games for the Astleys: 87 putouts, 191 assists, 10 errors, 30 double plays, .965 fielding percentage—278 total plays. Jody Reed has started 57 for Fugging: 74 putouts, 169 assists, 7 errors, 25 double plays, .972 fielding percentage—243 total plays. Reed's percentage is higher than Gagne's, but he's handled 35 fewer balls. Howard Johnson—yes, Chris Broyles is playing Howard Johnson at shortstop for the Wankers—has started 55 games there: 83 putouts, 120 assists, 2 errors, 19 double plays, .990 fielding percentage. His percentage is the best of any regular shortstop in the WLB. His 203 total chances are the fewest. By a wide margin. He has the cleanest hands in the league and the shortest reach.
Dickie Thon at short for Knockemstiff: 75 putouts, 143 assists, 9 errors, 30 double plays, .960 fielding percentage—the worst percentage of any regular shortstop. But his 227 total chances rank ahead of Johnson and Reed, meaning he's getting to balls that some others aren't. The fielding percentage says Thon is the worst defensive shortstop in the league. The total chances say he's providing more range value than Johnson and Reed. The gap between those two stories is the entire thesis of this column.
One of the most counterintuitive findings from the simulation research is which positions are most penalized by poor defenders—and the answer is not what conventional baseball wisdom would tell you.
Shortstop is important. But catcher and all three outfield positions are actually penalized more heavily than shortstop when substandard players man those spots. Third base is penalized more heavily than shortstop as well. The engine reflects a reality that the traditional defensive spectrum doesn't fully capture: the outfield positions involve enormous territory, and the difference between an Excellent center fielder and a Poor one is not a matter of a few extra grounders. It's a matter of balls falling for doubles and triples that an elite defender catches routinely.
The center field numbers are staggering. The community research found that an Excellent-range center fielder with a low error rate can save roughly 33 to 34 runs over a season compared to an Average/100 defender at the same position. Thirty-four runs. That's more than three wins from a single defensive position.
Look at what's happening in center field right now. Kirby Puckett has started 57 games in center for Nicaragua: 179 putouts, 2 assists, 1 error, .995 fielding percentage—182 total chances. Tony Gwynn has started 53 for Iron Knob: 128 putouts, 1 assist, 1 error, .992 fielding percentage—130 total chances. Robin Yount has started 58 in center for Huanca: 169 putouts, 2 assists, 3 errors, .983 fielding percentage—174 total chances. Ellis Burks and Roberto Kelly are splitting center for Fugging with a combined 178 putouts, 3 assists, 3 errors, .984 percentage. Paul O'Neill is playing center for Knockemstiff—a corner outfielder in center—with 138 putouts, 1 assist, and 5 errors in 46 starts, a .965 fielding percentage.
O'Neill's 5 errors are more than Puckett, Gwynn, and Yount combined. But the real damage isn't the errors. It's the 41-putout gap between Puckett's 179 and O'Neill's 138 in roughly the same number of starts. Those extra balls Puckett catches are balls that fall for hits against Knockemstiff. In a park with a 112 home run factor on both sides, a ball that falls in center field doesn't just become a single—it can rattle to the wall for a double, or score a runner from first. When your center fielder is a converted corner outfielder with 5 errors and the lowest putout total among contender center fielders, the pitching staff pays for it every night. And in Knockemstiff's case, the pitching staff is already the weakest in the field at 4.73 ERA. Brett Houlberg's 4.73 isn't just a pitching problem. It's a center field problem wearing a pitching disguise.
Ruben Sierra is also splitting time in center for Knockemstiff—12 starts, 36 putouts, 4 errors, a .900 fielding percentage. Nine combined errors from the center field position. The center of Knockemstiff's defense is hemorrhaging.
At first base, the traditional view holds that defense is almost irrelevant. The simulation is more nuanced. A first baseman with Excellent range extends the reach of his fellow infielders by handling throws in the dirt, scooping wide throws, and converting plays that a lesser first baseman lets go for errors charged to the shortstop or third baseman. Will Clark has started 54 games at first for Iron Knob: 553 putouts, 37 assists, 1 error, 60 double plays, .998 fielding percentage. One error in 591 total chances. Mark Grace for the Astleys: 331 putouts, 28 assists, 1 error, 27 double plays in 41 starts. Clark and Grace are the best first basemen in the league. Kent Hrbek for Huanca: 453 putouts, 29 assists, 5 errors, 43 double plays. Alvin Davis for Fugging: 439 putouts, 36 assists, 4 errors, 36 double plays. Fred McGriff for Knockemstiff: 329 putouts, 25 assists, 4 errors, 33 double plays. Don Mattingly for Nicaragua: 445 putouts, 26 assists, 4 errors, 32 double plays.
Clark's one error in 591 chances isn't just a first baseman being competent. It's a first baseman saving his fellow infielders—particularly Ozzie Smith—from errant throws that would otherwise become errors or extra bases. When you pair Excellent range at short with Excellent range at first, the left side of the infield becomes a wall. Iron Knob turns 64 double plays as a team—the most among contenders. Knockemstiff turns 38—the fewest. Clark and Smith are the foundation of that gap.
At third base, range matters more than most people think—and two contenders are paying a severe tax.
Bobby Bonilla has started 46 games at third for Knockemstiff: 32 putouts, 50 assists, 13 errors, 4 double plays, .863 fielding percentage. Thirteen errors. That's the most at any single position by any regular player on any contender. When Bonilla isn't there, Kevin Mitchell has filled in for 12 starts: 7 putouts, 14 assists, 3 errors, .875 fielding percentage. Combined, Knockemstiff's third basemen have committed 17 errors in 58 games—a number that demands its own paragraph, which it just received. Their .864 combined fielding percentage at third base is the worst at any position on any contender by more than sixty points.
But here's what the research tells us about Bonilla's situation: the 13 errors, while alarming, are not necessarily the primary cost. If Bonilla has strong range—and his 82 total chances in 46 starts suggests he's getting to some balls—the errors are partially compensated by the plays he makes that a lesser-ranged third baseman would let through for hits. An Excellent third baseman versus a Poor one is a gap of roughly 15 to 20 runs over a full season. Bonilla's error rate is severe. His range may be saving the Slap Daddies more than the errors are costing them. The fielding percentage screams disaster. The reality is more complicated—though "more complicated" does not mean "acceptable."
Now compare across the league. Terry Pendleton at third for Iron Knob: 39 putouts, 125 assists, 4 errors, 14 double plays, .976 fielding percentage—164 total chances in 54 starts. Pendleton has handled 82 more balls than Bonilla in 8 more starts. He's getting to twice as many plays and committing less than a third of the errors. Kelly Gruber at third for Huanca: 32 putouts, 98 assists, 10 errors, .929 fielding percentage—130 total chances in 58 starts. Gruber has 10 errors—second-most among contender third basemen—but 48 more total chances than Bonilla in roughly the same number of starts. The research tells us that those 48 extra plays represent substantial defensive value that the fielding percentage doesn't capture.
Carney Lansford at third for Fugging: 46 putouts, 94 assists, 7 errors, .952—140 total chances in 57 starts. Kevin Seitzer for Nicaragua: 16 putouts, 64 assists, 6 errors, .930—80 total chances in 48 starts, the fewest of any contender third baseman. Wade Boggs for the Astleys: 33 putouts, 62 assists, 3 errors, .969—95 total chances in 46 starts.
Pendleton is the gold standard. Seitzer and Boggs have the lowest total chances—Seitzer's 80 in 48 starts is remarkably few. The simulation values range above all else. At third base, the gap between Pendleton's 164 chances and Seitzer's 80 is 84 plays. At 15 to 20 runs separating Excellent from Poor over a full season, the defensive spectrum at the hot corner is not a footnote. It is a structural element of the pennant race.
Ryne Sandberg at second for Huanca: 108 putouts, 191 assists, 1 error, 39 double plays, .997 fielding percentage in 58 starts—299 total chances. One error. In the entire season. That is the best defensive performance at any position on any contender by a significant margin.
Steve Sax for Iron Knob: 121 putouts, 188 assists, 4 errors, 47 double plays in 55 starts—309 total chances, .987 fielding percentage. Sax has handled 10 more total chances than Sandberg in 3 fewer starts and turned 8 more double plays, suggesting comparable or slightly better range with marginally worse hands.
Paul Molitor for Knockemstiff: 113 putouts, 132 assists, 2 errors, 29 double plays in 58 starts—245 total chances, .992 fielding percentage. Clean hands—but 54 fewer total chances than Sandberg in the same number of starts. That 54-play gap is range, and it translates directly into extra baserunners against Knockemstiff's pitching staff. Lou Whitaker for the Astleys: 63 putouts, 77 assists, 2 errors, 16 double plays in 33 starts—140 total chances (with Robby Thompson and Jeff Treadway splitting the remaining time).
Roberto Alomar for Nicaragua: 80 putouts, 133 assists, 10 errors, 23 double plays in 48 starts—213 total chances, .955 fielding percentage. That's the worst fielding percentage among contender second basemen by a wide margin. Alomar—one of the most celebrated defensive players of his era—has committed 10 errors at second, more than Sandberg, Molitor, and Sax combined. The research tells us not to overreact—if Alomar's range is strong, the 10 errors are partially compensated by the plays he's making that others wouldn't reach. But his 213 total chances in 48 starts (4.4 per start) compare unfavorably to Sandberg's 299 in 58 starts (5.2 per start). Sandberg is getting to more balls and making virtually no errors. Alomar is getting to fewer balls and making more errors. That combination is part of the explanation for why Nicaragua's IPAVG sits at .289 despite owning the league's best K/BB ratio at 3.2.
Julio Franco for Fugging: 75 putouts, 126 assists, 6 errors, 26 double plays in 48 starts—201 total chances, .971 fielding percentage. Franco's range total is the lowest per start among contender second basemen, and his 6 errors compound the issue.
Outfielder throwing ratings operate on a plane separate from range. Range determines whether the ball is caught. Throwing determines what happens after the ball is fielded—or, more precisely, what happens to the runners while it's being fielded. An outfielder with an Excellent throwing arm doesn't just gun down runners. The arm's primary value is deterrence: runners don't take the extra base. These non-events—the bases not taken, the runs not scored—are invisible in the box score but fully modeled in the simulation. Over a full season, the gap between an Excellent-arm right fielder and a Poor-arm right fielder is not merely a handful of assists. It's dozens of instances where runners hold at second instead of advancing to third—all of which reduces the number of runners in scoring position and suppresses the opponent's ability to manufacture runs.
The team stolen base numbers illuminate the catcher side. Huanca's catchers—Brian Harper primarily, starting 51 games—have allowed only 28 stolen bases while throwing out 25, a .528 stolen base percentage that is the lowest among contenders. Iron Knob's catchers have allowed 14 steals and caught 8, for a .636 rate—but the raw volume is the story: only 14 stolen base attempts in 55 games means opponents aren't running. That's deterrence. The Astleys allow 25 steals against 11 caught (.694). Fugging allows 29 against 12 (.707). Knockemstiff allows 28 against 8 (.778)—the highest rate among contenders. Nicaragua allows 22 against 17 (.564), with Terry Steinbach's arm doing the heavy lifting.
The gap between Huanca's .528 and Knockemstiff's .778 is the difference between a running game that opponents think twice about and a running game that operates freely. Over 154 games, that gap translates to additional bases, additional runs, and additional losses—a tax that compounds silently.
Now pull back and look at the aggregate team fielding through June 6, layered against the team pitching and the head-to-head records, and several things crystallize.
The Rick Astleys lead the WLB with a .986 team fielding percentage—31 errors in 60 games. Their pitching staff has a 3.56 ERA and a 1.16 WHIP, the lowest in the league. Their in-play batting average against is .249—also the lowest. The defense is converting batted balls into outs at the highest rate of any contender, and the pitching staff's component numbers reflect it. Their RCERA of 3.35 is actually lower than their actual ERA of 3.56—a .21 gap suggesting the pitching staff is slightly unlucky, partly because of the 123-percent left-handed home run park factor I discussed in my ballpark column. The .986 fielding percentage and .249 IPAVG confirm the defense is doing its job on balls in play. The home runs—73 allowed, the most among contenders—are the problem. The defense is the answer to everything else.
Iron Knob sits at .985—one point behind the Astleys—with 32 errors in 55 games. But the assists tell a different story entirely. Iron Knob has 667 assists as a team—101 more than the Astleys' 566, and 218 more than Knockemstiff's 449. Assists measure plays made on the run—groundballs fielded and thrown, relay throws, double plays turned. Iron Knob's 64 double plays lead the league among contenders. The Astleys have 43. Knockemstiff has 38. The assists-per-game rate measures how many batted balls the defense is converting into outs through active play-making: 12.1 assists per game for Iron Knob versus 9.4 for the Astleys, 9.4 for the Wankers, and 7.7 for the Slap Daddies. That 4.4-assist-per-game gap between Iron Knob and Knockemstiff is the glove tax, expressed in its purest form.
And the pitching numbers confirm the mechanism. Iron Knob's ERA is 3.46 against a RCERA of 3.45—a gap of .01. I've written repeatedly that this is the most trustworthy ERA in the league. Now you can see why. The defense is converting the pitchers' batted balls into outs at precisely the rate the component stats predict. No leaks. No extra singles bleeding through the left side. No doubles in the gap that should have been caught. When Hershiser induces weak contact, Smith and Pendleton and Clark and Sax convert it into outs with mechanical precision. The .01 ERA-to-RCERA gap I've been celebrating for three columns isn't just a pitching number. It's a pitching-and-defense number. The two are inseparable.
Knockemstiff sits at .974 fielding percentage—the worst of any contender. Fifty-three errors in 58 games. Their pitching staff's ERA is 4.73, with a RCERA of 4.72—a gap of .01, just like Iron Knob. What you see is what you get. But what you get is 1.27 ERA runs worse. The 4.73 ERA is the true talent level of Knockemstiff's pitching-and-defense unit—and that unit includes a defense that commits the most errors among contenders, posts the lowest assists total (449), and turns the fewest double plays (38). When anyone other than Nolan Ryan pitches—Ryan's 12.1 K/9 bypasses the defense by striking batters out—the balls in play find holes because the defense can't close them.
Fugging is at .982 with 37 errors. Huanca is at .981 with 40. Nicaragua is at .977 with 49 errors—the second-worst among contenders, and this number demands scrutiny.
Nicaragua's 49 errors are the most damaging surprise in this analysis. Alomar's 10 at second, Seitzer's 6 at third, Saberhagen's 3 on the mound, Bob Welch's 3 on the mound—the errors are distributed broadly rather than concentrated, which makes them harder to fix. Nicaragua's pitching staff carries the league's best K/BB ratio at 3.2, generating strikeouts and limiting walks at a rate that should minimize the defense's exposure. Yet their IPAVG is .289—tied with Knockemstiff and Fugging for the second-highest cluster among contenders. With the league's best command profile, Nicaragua's pitchers should be posting a lower IPAVG. That they aren't points to a defensive structure that is letting balls through at a rate that partially offsets the pitching staff's excellence. The gap between Nicaragua's ERA (3.85) and RCERA (3.74)—the pitching staff performing .11 runs worse than its components predict—may have more to do with the 49 errors and the range limitations at second and third than with any failing on the mound.
Every offseason, every trade deadline, the WLB's owners face the same question: where do I invest? The answer, for most, has been pitching and offense. Draft the ace. Trade for the slugger. Stack the lineup with OBP machines and power bats. Defense has been treated as a luxury—nice to have, not essential to contend.
The simulation says otherwise. When a single position—center field, shortstop, catcher—can swing twenty to thirty-five runs over a season based solely on the range and throwing ratings of the player manning it, defense is not a luxury. It is infrastructure. It is the foundation on which pitching performance is built. A groundball pitcher without a good shortstop is a pitcher whose groundball tendency is partially wasted. A flyball pitcher without a good center fielder is a pitcher whose fly ball outs become gap doubles. The pitching ratings set the ceiling. The defensive ratings determine how close the team gets to that ceiling.
The research tells us that an owner should never start a Poor-range defender at a premium position if any alternative exists—the run cost is too steep. It tells us that error rate, while not irrelevant, is dramatically less important than range—you should take the player who gets to more balls and bobbles some over the player who only reaches the routine ones and fields them all cleanly. It tells us that outfield arms matter more than the traditional defensive spectrum suggests—the bases not taken are the runs not scored, and they compound silently across 154 games. It tells us that a first baseman with Excellent range makes every infielder around him better by saving their errant throws, an effect that shows up in team fielding percentage but is actually a product of the first baseman's skill.
And it tells us something about fielding average—that sacred old statistic—that we should probably have understood long ago: it measures the wrong thing. Fielding average tells you how often a player handles the balls he reaches without making an error. It tells you nothing about the balls he never reached. Howard Johnson's .990 fielding percentage at short for Huanca is the best in the league. His 203 total chances are the fewest. He has the cleanest hands and the shortest reach. Ozzie Smith's .975 is fifteen points worse. His 307 total chances are the most. The .975 shortstop is dramatically more valuable. The engine knows it. The standings reflect it.
Andrew Harris built Iron Knob the way the simulation rewards. Elite pitching behind elite defense. Ozzie Smith at short. Terry Pendleton at third. Will Clark at first. Steve Sax at second. Tony Gwynn in center. This is not a roster built by accident. It is a roster built on understanding that in Diamond Mind, roughly seventy percent of all outs are recorded by fielders, and the difference between an Excellent fielder and a Poor one at a premium position is worth more than the gap between a good starter and a mediocre one. The 37-18 record, the 25-9 road mark, the .01 ERA-to-RCERA gap, the 667 assists, the 64 double plays—all of it flows from the same principle. Elite defense is the foundation. Elite pitching is the structure built on top of it.
The glove tax is real. Some teams are paying it. Some teams are collecting it. The standings are sorting accordingly.
Iron Knob has 667 assists and 64 double plays—the most of any contender. Knockemstiff has 449 assists and 38 double plays—the fewest. The gap in ERA between the two staffs is 1.27 runs. The gap in defense is where that 1.27 lives.