Chapter 17 Quiz: Modeling MLB
Test your understanding of MLB-specific modeling concepts, sabermetric foundations, and betting market patterns.
Question 1. What does wOBA stand for, and why is it considered more useful for betting models than traditional batting average?
Answer
wOBA stands for weighted On-Base Average. It is more useful than batting average because it assigns empirically derived linear weights to each plate appearance outcome (walk, single, double, triple, home run) based on their actual run value. Batting average treats all hits equally and ignores walks entirely, which means it fails to capture the true offensive contribution of a hitter. For betting models, wOBA's stronger correlation with run scoring makes it a superior input for projecting game-level run production.Question 2. What is FIP, and why is it more predictive of future pitcher performance than ERA?
Answer
FIP is Fielding Independent Pitching. It measures pitcher quality using only the three outcomes a pitcher directly controls: strikeouts, walks (plus hit-by-pitches), and home runs. ERA, by contrast, includes the effect of defensive quality, sequencing luck, and BABIP variance. Because these factors are largely outside the pitcher's control and are highly variable, FIP strips them away to isolate the pitcher's true skill. Empirically, FIP has a higher year-over-year correlation than ERA, making it the better predictor of future run prevention.Question 3. What is the approximate stabilization threshold for strikeout rate (K%) versus BABIP for hitters?
Answer
Strikeout rate stabilizes at approximately 60 plate appearances, making it one of the fastest-stabilizing batting metrics. BABIP (batting average on balls in play) requires approximately 800 or more plate appearances to stabilize, which exceeds most hitters' full-season totals. This means K% is trustworthy very early in the season, while BABIP-dependent metrics like batting average require heavy regression to the mean for months.Question 4. A pitcher has an ERA of 2.50 and a FIP of 4.10. What does this divergence most likely indicate?
Answer
This divergence most likely indicates that the pitcher has been extremely lucky. His defense has probably converted an unusually high percentage of balls in play into outs (low BABIP against), and/or he has stranded an unusually high percentage of baserunners (high LOB%). His true talent level for run prevention is much closer to the 4.10 FIP, and his ERA should be expected to regress sharply upward. This creates a betting opportunity: fade this pitcher's inflated reputation by betting against his team or on overs in his starts.Question 5. What is the Pythagorean expectation, and what exponent is typically used for MLB?
Answer
The Pythagorean expectation estimates a team's expected win percentage from its runs scored and runs allowed: $\text{Win\%} = \frac{\text{RS}^{k}}{\text{RS}^{k} + \text{RA}^{k}}$. For MLB, the empirically derived exponent is approximately 1.83 (sometimes 2 is used as a simpler approximation). Teams whose actual record significantly exceeds or falls short of their Pythagorean expectation are expected to regress, making this a useful tool for identifying overvalued and undervalued teams.Question 6. Explain the concept of "platoon splits" in baseball. Which matchup combination generally produces the lowest wOBA for hitters?
Answer
Platoon splits refer to the systematic performance differences that arise from the handedness matchup between pitcher and batter. Hitters generally perform better against opposite-handed pitchers and worse against same-handed pitchers. The matchup that produces the lowest wOBA for hitters is left-handed batters facing left-handed pitchers (LHB vs. LHP), with an approximate wOBA of .305, compared to the most favorable matchup of right-handed batters facing left-handed pitchers (RHB vs. LHP) at approximately .335.Question 7. What is the approximate run park factor for Coors Field, and what are the three primary mechanisms by which altitude affects run scoring?
Answer
Coors Field has a run park factor of approximately 1.35 (35% above neutral). The three primary mechanisms are: (1) reduced air density at 5,280 feet reduces drag on the baseball, causing batted balls to travel 5-9% farther than at sea level; (2) reduced air density causes breaking pitches to break less, making them easier to hit; and (3) fastballs, while not losing velocity, have less "life" or movement due to reduced Magnus effect, slightly reducing their effectiveness.Question 8. Why does the negative binomial distribution fit MLB run-scoring data better than the Poisson distribution?
Answer
The negative binomial fits better because MLB run scoring exhibits overdispersion: the variance of runs per game exceeds the mean, violating the Poisson assumption that variance equals the mean. The overdispersion arises from run clustering in baseball, where big innings (multiple runs in a single inning) occur more frequently than a Poisson process would predict. The negative binomial has an extra parameter (dispersion parameter $r$) that captures this excess variance. Empirically, $r \approx 5$--$8$ fits MLB data well.Question 9. What is the standard MLB run line, and why does the relationship between moneyline probability and run line probability create betting value?
Answer
The standard MLB run line is $\pm 1.5$ runs. The relationship creates value because it is non-linear: a team with a 67% moneyline win probability ($-200$ favorite) covers $-1.5$ only about 55% of the time, because many wins come by exactly 1 run. This non-linearity means that: (a) heavy favorites on the run line often offer better relative value than the moneyline because the implied probability gap is larger, and (b) heavy underdogs on $+1.5$ are frequently overpriced relative to their true cover probability.Question 10. What are "first five innings" (F5) bets, and why are they attractive for bettors with strong pitcher models?
Answer
F5 bets settle based on the score after 5 complete innings, effectively removing bullpen variance from the equation. They are attractive for pitcher-model bettors because: (1) the starting pitcher typically covers most or all of the first 5 innings, making the bet a purer test of the starting pitcher matchup model; (2) bullpen performance is harder to predict day-to-day; (3) F5 totals are lower (typically 4--5 runs), where Poisson-based models are more accurate; and (4) F5 markets may be less efficiently priced than full-game markets.Question 11. What is Stuff+ and how is it used in pitcher evaluation?
Answer
Stuff+ is a composite pitch-quality metric derived from Statcast data that evaluates individual pitches based on their physical characteristics (velocity, movement, spin, release point) rather than their outcomes. A Stuff+ of 100 is league average, with higher values indicating better pitch quality. It is useful for pitcher evaluation because it measures the process (how good are the pitches?) rather than the outcome (what happened after the pitch?), making it a leading indicator that can identify pitchers whose results have not yet caught up to their pitch quality.Question 12. Explain reverse line movement (RLM) in MLB betting. Why is it particularly informative in baseball?
Answer
Reverse line movement occurs when the betting line moves in the opposite direction from the public betting percentages. For example, if 70% of bets are on Team A, but the line moves to make Team A less favored, it suggests large sharp bets are on Team B. RLM is particularly informative in MLB because: (1) the moneyline structure allows sharp money to be camouflaged in underdog prices; (2) the high volume of games (2,430 per season) provides many opportunities; and (3) late-breaking information such as weather, lineups, and bullpen availability creates windows where sharp bettors have informational edges.Question 13. How do umpire effects influence MLB totals betting, and what is the approximate magnitude of the effect?
Answer
Home plate umpires influence totals through their strike zone tendencies. Umpires with large strike zones generate more called strikes, suppress offense, and lead to lower-scoring games. Umpires with small zones do the opposite. The effect ranges from approximately 0.5 to 1.0 runs per game between the most generous and most restrictive umpires. For a game with a posted total of 8.5, a 0.5-run shift changes the over/under probability by approximately 5-8 percentage points, which is a significant edge.Question 14. A batter has an actual wOBA of .290 and an xwOBA (from Statcast) of .345. What does this gap imply?
Answer
This gap implies the batter has been significantly unlucky. His quality of contact (as measured by exit velocity and launch angle) suggests he "should" be producing at a .345 wOBA level, but the actual outcomes have been worse. The most likely cause is bad luck on batted ball placement (BABIP below expected). The xwOBA-to-wOBA gap signals strong regression potential: the batter's future production is likely to be much closer to .345, making his team potentially undervalued by the market if bettors are anchoring on the actual .290.Question 15. What is the bullpen leverage index (LI) and how is it used in bullpen modeling for betting?
Answer
Leverage Index measures the importance of a game situation based on the inning, score, and baserunners. A leverage index of 1.0 is average; high-leverage situations (close games, late innings, runners on base) have LI values above 2.0. For bullpen modeling, LI is used to weight reliever contributions because managers deploy their best relievers in high-leverage situations. The bullpen rating formula uses innings-and-leverage-weighted FIP: $\text{Bullpen Rating} = \frac{\sum \text{IP}_i \times \text{LI}_i \times \text{FIP}_i}{\sum \text{IP}_i \times \text{LI}_i}$, giving more weight to the relievers who pitch in the situations that matter most.Question 16. What is xFIP and how does it differ from standard FIP?
Answer
xFIP (Expected Fielding Independent Pitching) modifies FIP by replacing a pitcher's actual home run total with an expected home run total based on fly ball rate and the league-average HR/FB rate. This further removes variance from the metric, since home run rates on fly balls fluctuate significantly for individual pitchers from year to year. xFIP is slightly more predictive of future ERA than FIP, particularly for pitchers with extreme HR/FB rates that are likely to regress.Question 17. Why should small-sample batter-vs-pitcher historical matchup data (e.g., "7-for-12 lifetime") be ignored in betting models?
Answer
Small-sample head-to-head records are almost entirely noise. With only 12 plate appearances, the variance is enormous: a true-.300 hitter could easily go 7-for-12 or 2-for-12 by chance alone. The binomial confidence interval for 7/12 (.583) ranges from roughly .277 to .849, which is essentially uninformative. The predictive power of 12 PAs is negligible compared to skill-level reasoning (e.g., "this hitter struggles against high-velocity fastballs"). Basing bets on these records introduces noise rather than signal into the model.Question 18. How does the wind affect MLB scoring, and what is the approximate magnitude of a 15 mph wind blowing out?
Answer
Wind is the most impactful single-game weather variable in MLB. A 15 mph wind blowing out to center field can add 20-30 feet to fly ball distance, dramatically increasing home run probability and overall scoring. Using the environmental adjustment coefficients from the chapter ($\text{WIND\_OUT\_COEFF} = 0.008$ per mph), a 15 mph wind adds approximately $15 \times 0.008 = 12\%$ to the park factor. For a game projected at 8.5 total runs, this adds approximately 1.0 run, potentially pushing the expected total to 9.5 or higher.Question 19. What does wRC+ of 130 mean, and why is it preferred for cross-park comparisons?
Answer
A wRC+ of 130 means the hitter produced 30% more runs per plate appearance than the league average, adjusted for both park effects and the run-scoring environment of the league. wRC+ is preferred for cross-park comparisons because it normalizes away the park factor: a hitter with 130 wRC+ at Coors Field and a hitter with 130 wRC+ at Oracle Park are evaluated as equally productive relative to their contexts. Without park adjustment, Coors hitters would appear systematically better than they truly are.Question 20. Explain the "April effect" in MLB betting and how it creates opportunities for informed bettors.
Answer
In April, uncertainty is at its highest: preseason projections are imprecise, current-year sample sizes are tiny, and weather in northern cities introduces extra scoring variance. The market tends to overreact to hot and cold starts because small-sample batting averages and ERAs dominate the public narrative. Informed bettors can exploit this by: (1) trusting fast-stabilizing metrics (K%, exit velocity) over slow ones (batting average, ERA); (2) identifying regression candidates whose underlying contact quality or pitch quality diverges from their surface stats; and (3) recognizing that cold, wet April weather suppresses scoring more than the market sometimes prices.Question 21. What is SIERA and what additional information does it capture beyond FIP?
Answer
SIERA (Skill-Interactive ERA) is an advanced pitching metric that accounts for the interaction between strikeout rate and ground ball rate. Unlike FIP, which treats K%, BB%, and HR independently, SIERA recognizes that a pitcher who strikes out a lot of batters AND generates ground balls is more effective than the sum of those skills would suggest (because the ground balls are more likely to be outs when fewer balls are put in play). SIERA also accounts for the tendency of fly-ball pitchers to allow more home runs. It is generally considered the most predictive pitching metric for future ERA.Question 22. Why are MLB betting markets considered among the most efficient in sports for moneylines with known starting pitchers?
Answer
MLB moneyline markets with known starters are highly efficient because: (1) baseball's deep statistical tradition provides rich, publicly available data; (2) the starting pitcher matchup, the dominant factor in game outcomes, is announced in advance; (3) the 162-game season provides massive sample sizes for model calibration; (4) many sophisticated quantitative bettors (sharp money) compete in this market; and (5) sportsbooks have long experience pricing baseball games. Finding edges requires integrating information beyond what FanGraphs or Baseball Reference provide, such as real-time weather, bullpen fatigue, and lineup composition.Question 23. How would you use the Odds Ratio method to estimate a batter's expected wOBA against a specific pitcher?
Answer
The Odds Ratio method combines the batter's skill, the pitcher's quality, and the platoon matchup by computing: $\text{Expected wOBA} = \text{league wOBA} \times \frac{\text{batter wOBA}}{\text{league wOBA}} \times \text{pitcher factor} \times \text{platoon factor}$. The batter ratio is their wOBA divided by league average. The pitcher factor converts the pitcher's quality score to a wOBA-against multiplier. The platoon factor adjusts for the handedness matchup. This multiplicative approach properly accounts for the fact that a great hitter facing a great pitcher does not produce average results; instead, both effects partially offset.Question 24. What is the trade-deadline effect in MLB betting, and how long does the market typically take to fully price roster changes?
Answer
The trade deadline (late July/early August) creates sharp changes in team composition that can shift a team's projected runs per game by 0.2-0.5 in either direction. The market typically takes 1-2 weeks to fully price these changes. During this window, the acquiring team may be undervalued (the market has not yet fully credited the upgrade) and the selling team overvalued (the market has not yet fully penalized the loss of talent). Bettors who immediately re-run their models with updated rosters can capture value during this adjustment period.Question 25. A team has scored 400 runs and allowed 380 runs through 80 games. Their actual record is 46-34 (.575). Calculate their Pythagorean win percentage and determine whether they are over- or underperforming. What does this imply for their second-half betting value?