A comprehensive educational reference covering every analytical framework used in professional baseball handicapping, from pitching metrics and park factors to market structure and bankroll discipline.
This guide is designed as a central reference for anyone looking to build a structured, analytically grounded approach to evaluating Major League Baseball. It is not a shortcut, a system, or a formula. It is a framework. Each section covers one analytical pillar of baseball handicapping: the metrics that matter, why they matter, and how to apply them within a repeatable evaluation process.
The content here is organized to be read in any order. Each section stands on its own as a reference for a specific analytical domain, from starting pitcher evaluation to bankroll management. At the same time, the sections build on each other. A comprehensive game evaluation draws from multiple pillars simultaneously: the starting pitching matchup, the park environment, the bullpen state, the lineup construction, the situational context, and the market itself. Reading this guide from start to finish will illustrate how those pillars connect.
Throughout this guide, links to dedicated deep-dive pages are provided for each topic. Those pages expand significantly on the foundations outlined here, with detailed examples, data tables, and extended explanations. Think of this page as the table of contents and orientation layer, and the linked pages as the full chapters.
There are no guarantees embedded in any of this material. Baseball is a probabilistic game, and even the most rigorous process will produce losing days, losing weeks, and losing months. The objective is to develop a process that identifies value consistently across a large sample of evaluations. This guide is the educational foundation for building that process.
Earned Run Average is the most widely cited pitching statistic in baseball, and it is also one of the least reliable for projecting future performance. ERA measures results, but it does not isolate the pitcher's contribution from the defense behind him, the ballpark he pitches in, or the sequencing luck inherent in when hits happen to fall. A pitcher who allows three singles in an inning gives up runs; a pitcher who allows those same three singles spread across three innings may not. ERA treats those two outcomes differently, but the pitcher's actual performance was identical.
FIP strips away the variables a pitcher cannot control, specifically balls in play, and focuses only on the three outcomes a pitcher directly governs: strikeouts, walks, and home runs. FIP is scaled to look like ERA for ease of comparison. A pitcher with a 3.20 ERA but a 4.10 FIP has likely been aided by strong defense or favorable sequencing. A pitcher with a 4.50 ERA but a 3.40 FIP has likely been hurt by bad luck on balls in play. Over time, FIP is a significantly better predictor of future ERA than ERA itself.
xFIP takes FIP one step further by normalizing home run rate. Instead of using the pitcher's actual home runs allowed, xFIP substitutes the league-average home run rate per fly ball. This is useful because home run rates fluctuate year-to-year more than most people realize. A pitcher who allows a high number of home runs in a small sample may simply be experiencing variance, not a fundamental flaw. xFIP smooths that out and provides a more stable baseline.
SIERA is the most complex of the common pitching estimators and arguably the most predictive. It accounts for the relationship between strikeout rate, walk rate, and ground ball rate, recognizing that these skills interact in non-linear ways. A high-strikeout, high-ground-ball pitcher is more effective than the sum of those two skills would suggest, because strikeouts prevent the ball from going into play and ground balls limit damage when contact occurs. SIERA captures this synergy. For game-level evaluation, SIERA provides the most reliable baseline projection of a pitcher's true run-prevention ability.
The strikeout-to-walk ratio (K/BB) is one of the oldest and most enduring indicators of pitching quality. A K/BB above 3.0 is strong; above 4.0 is elite. This ratio matters because it reflects command, the ability to miss bats while also avoiding free bases. Called Strikes plus Whiffs percentage (CSW%) is a more granular measure that captures how often a pitcher generates desirable outcomes on a per-pitch basis. A CSW% above 30% indicates a pitcher who controls at-bats. Below 26% signals a pitcher who is working uphill on most counts.
When evaluating a game-day starting pitching matchup, comparing xFIP, SIERA, K/BB, and CSW% between the two starters provides a strong analytical foundation. These metrics are more predictive than ERA, win-loss record, or quality start percentage.
When metrics conflict, prioritize in this order: SIERA (most predictive), xFIP (second most stable), FIP (accounts for actual HR allowed), ERA (most influenced by external factors). Use ERA only as context for how results have materialized, not as a projection tool.
For an expanded breakdown of pitching evaluation with seasonal data tables and per-start application examples, see the Starting Pitcher Analysis page.
Modern bullpen usage has fundamentally changed how baseball games play out. Starting pitchers now average fewer than five and a half innings per start across the league, which means three and a half to four and a half innings, roughly 40% of the game, are pitched by relievers. Any evaluation that focuses only on the starting matchup and ignores the bullpen is leaving out nearly half the game.
Not all relief innings are equal. Leverage Index (LI) quantifies how important a given plate appearance is relative to the game's outcome. An LI of 1.0 is average. An LI above 2.0 is a high-leverage situation where the outcome of the at-bat has an outsized impact on win probability. The best bullpens have elite arms deployed specifically in these high-leverage situations, often in the seventh and eighth innings rather than saving them exclusively for the ninth. Understanding which teams deploy high-leverage relievers flexibly versus rigidly affects how likely a lead is to hold.
Reliever effectiveness degrades with workload. Pitch counts over the previous three days are the most commonly used fatigue indicator. A reliever who has thrown 40 or more pitches in the preceding 72 hours is significantly less effective than his season-long metrics suggest. This is especially relevant during series where the same bullpen arms may have been deployed on consecutive nights. Checking a bullpen's recent workload, not just its seasonal numbers, is essential for late-game evaluation.
Platoon advantages are magnified in bullpen situations because managers actively construct favorable handedness matchups. Left-handed batters hit roughly 30 points lower in batting average against left-handed relievers than against right-handed relievers. When a team has left-handed specialist options available and the opposing lineup is heavy with left-handed bats, the bullpen advantage is compounded. Evaluating which specialists are available on a given night is part of a thorough game-day process.
Bullpen depth, availability, and deployment patterns introduce variance that pure starting-pitcher analysis cannot capture. A strong starter handing the ball to a depleted bullpen is a fundamentally different proposition than handing it to a fresh, well-rested one. For a detailed methodology on bullpen evaluation and fatigue tracking, see the Bullpen Analysis Guide.
Always check the bullpen usage log before evaluating late-game scenarios. Sites like FanGraphs and Baseball Savant publish daily pitch count logs for every reliever. A bullpen that looks elite on paper may be functionally compromised if its best arms threw 30+ pitches the night before.
Park factors quantify how a stadium's physical environment affects offensive production relative to the league average. They are expressed as an index where 100 represents a neutral environment. A run index of 110 means games in that park produce roughly 10% more runs than average. A run index of 90 means 10% fewer. Park factors are calculated over multi-year samples, typically three to five seasons, to smooth out seasonal anomalies.
There are separate park factor indices for runs, home runs, hits, doubles, triples, and strikeouts. The run index is the most broadly applicable, but the home run index is particularly useful for evaluating power-dependent outcomes. A park can suppress overall run scoring while still being above average for home runs (American Family Field in Milwaukee is a historical example), or it can boost overall offense while suppressing home runs specifically (Fenway Park, where the Green Monster generates doubles rather than clearing the fence).
Coors Field in Denver sits at 5,280 feet above sea level, where air density is approximately 18% lower than at sea level. This affects baseball in three ways: fly balls carry farther due to reduced drag, breaking pitches lose movement because the thinner air provides less surface interaction, and the overall offensive environment is amplified to the point where Coors Field's multi-year run index typically sits between 115 and 130. No other park in baseball comes close to this magnitude of offensive inflation.
Beyond atmospheric conditions, the geometry of the playing field matters. Short outfield fences and low wall heights produce more home runs. Deep alleys and high walls produce more extra-base hits but fewer home runs. Oracle Park in San Francisco has one of the deepest right-center field dimensions in baseball, paired with a marine layer that suppresses ball flight, producing a consistently low home run index. Great American Ball Park in Cincinnati has short fences and low walls, producing one of the highest home run indices in the league.
Understanding park factors is foundational for game-day totals evaluation. A pitching matchup that projects at 7.5 runs in a neutral environment may project at 9 runs at Coors Field and 6.5 runs at T-Mobile Park. For a complete stadium-by-stadium breakdown with data tables and application methodology, see the Park Factors Guide.
Static park factors provide a baseline, but daily weather conditions can shift a park's run environment significantly in either direction. Temperature, humidity, wind speed, and wind direction all affect how far a baseball travels and how much movement a pitched ball generates. Ignoring weather is ignoring a variable that can shift expected run totals by one to two full runs in extreme cases.
Warmer air is less dense than cooler air, which means baseballs carry farther in hot weather. Research from Alan Nathan's physics of baseball models indicates that a fly ball hit in 95-degree weather travels approximately 6 to 8 feet farther than the same fly ball hit in 55-degree weather, all other variables being equal. In April games in northern cities, when temperatures may sit in the 40s and 50s, this suppression effect is substantial. In July, when temperatures are in the 80s and 90s, parks that are neutral by their static factor may play as hitter-friendly due to heat.
Contrary to popular belief, humid air is actually less dense than dry air at the same temperature, because water molecules are lighter than the nitrogen and oxygen molecules they displace. In theory, humid conditions should help ball flight. In practice, the effect is modest and often overwhelmed by other variables. The marine layer, which is common at West Coast parks like Oracle Park and Petco Park during evening games, is a separate phenomenon. The marine layer is a thick, cool air mass that rolls in from the ocean, creating a ceiling of dense air that physically suppresses ball flight. Night games at Oracle Park under a heavy marine layer can see fly ball distances drop by 10 to 15 feet compared to afternoon games at the same park.
Wrigley Field is the most wind-dependent park in baseball. When wind blows out toward the lake from the southwest, the park transforms into one of the most hitter-friendly environments in the league. When wind blows in from Lake Michigan, it becomes a pitcher's park. The difference between a strong wind-out day and a strong wind-in day at Wrigley can represent a swing of three to four expected runs. Other parks, particularly Kauffman Stadium and Guaranteed Rate Field, are also significantly wind-affected, though not to Wrigley's extreme.
Day games typically feature warmer temperatures and better visibility, which generally favors offense. Night games, especially on the West Coast, bring cooler air and in some cases marine layer effects. The day/night split is most pronounced at parks where the temperature differential is greatest, particularly in April, May, September, and October when daytime highs and nighttime lows can differ by 20 or more degrees.
For a comprehensive weather integration methodology, including wind speed thresholds and temperature adjustment scales, see the Weather Impact on MLB Betting page.
A starting lineup is not a fixed entity. Managers routinely adjust their lineups based on the opposing starter's handedness, the park, rest considerations, and matchup data. Evaluating a game based on a team's "usual" lineup without checking the actual posted lineup is a common source of error. The difference between a team's strongest possible lineup and its rest-day configuration can be significant, sometimes the equivalent of a full run of expected offense.
Weighted Runs Created Plus (wRC+) is the single best metric for evaluating how productive a lineup is relative to the league. It accounts for park factors and scales to a league average of 100. A lineup slot filled by a hitter with a wRC+ of 140 is producing at 40% above league average. A slot filled by a hitter with a wRC+ of 80 is 20% below. By adding up the wRC+ values for each lineup position and comparing the two teams' totals, you get a quick, park-adjusted read on the offensive talent gap. This is more informative than batting average, OPS, or runs per game.
Platoon splits are among the most persistent and well-documented patterns in baseball. On average across the league, left-handed batters perform significantly worse against left-handed pitchers than against right-handed pitchers. The reverse is also true, though typically to a lesser degree: right-handed batters have a smaller but measurable advantage against left-handed pitching. The magnitude of the platoon split varies by individual hitter. Some left-handed batters have a reverse platoon split (they hit lefties better), but those are the exception.
When a left-handed starting pitcher faces a lineup loaded with left-handed bats, the aggregate offensive expectation for that lineup decreases. When that same lineup faces a right-handed starter, the expectation increases. Evaluating how each lineup stacks up against the specific opposing starter's handedness is a step that pure team-level statistics miss entirely.
Bench quality matters more than most analyses account for. In close games, managers make pinch-hitting substitutions in the sixth, seventh, and eighth innings that can swing outcomes. A team with a deep bench full of platoon-advantaged options has a structural late-game advantage over a team with a thin bench. This is particularly relevant in National League-rules games and in interleague play where roster construction philosophies differ.
For detailed platoon split data and lineup evaluation frameworks, see the Platoon Splits Guide.
Offense in baseball is ultimately about two things: how often a hitter makes contact, and how hard that contact is. The modern analytical framework has moved beyond batting average and home runs to focus on process-oriented metrics that measure the quality of a hitter's approach and the quality of his contact. These metrics are more predictive of future performance than traditional results-based statistics.
Chase rate measures how often a batter swings at pitches outside the strike zone. The league average chase rate hovers around 28%. Hitters with chase rates above 33% are consistently vulnerable to pitchers who can locate off-speed pitches on the edges. Hitters with chase rates below 24% force pitchers into the zone, leading to more hittable pitches. At the team level, lineup-wide chase rate is one of the strongest predictors of offensive consistency. Teams that chase excessively are prone to high-strikeout, low-scoring outputs, especially against pitchers with elite off-speed stuff.
Statcast defines a "barrel" as a batted ball with a combination of exit velocity and launch angle that historically produces a batting average above .500 and a slugging percentage above 1.500. Barrel rate measures how often a hitter produces these optimal contact events. Hard-hit percentage measures the frequency of batted balls with exit velocities of 95 mph or higher. Both metrics are strong indicators of true offensive quality because they measure the process (contact quality) rather than the result (whether the ball happened to fall for a hit). A hitter with a high barrel rate and hard-hit percentage who is running a low batting average is likely underperforming due to sequencing or defensive alignment, not due to a decline in skill.
Batting Average on Balls in Play (BABIP) measures how often balls that are put in play (excluding home runs) fall for hits. The league average BABIP typically sits around .300. Individual hitters have stable long-term BABIPs based on their speed, contact quality, and batted ball distribution, but over stretches of 50 to 100 at-bats, BABIP fluctuates wildly due to luck. A hitter carrying a .380 BABIP over a three-week stretch is almost certainly benefiting from favorable sequencing and will regress. A hitter carrying a .220 BABIP over that same stretch is likely underperforming his true talent and will bounce back. Identifying BABIP-driven hot streaks and cold streaks is one of the most reliable ways to project offensive performance shifts.
When a hitter's BABIP deviates more than 40 points from his career average over a sample of 100+ plate appearances, regression is likely. Combine this with barrel rate and hard-hit rate: if those underlying quality metrics remain stable while BABIP is depressed, the hitter's batting average is expected to rise. If BABIP is elevated but barrel rate has declined, the batting average is expected to fall.
Baseball's 162-game schedule creates fatigue patterns, travel burdens, and motivational contexts that are measurable and analytically relevant. These situational factors operate independently of talent and can shift expected outcomes by meaningful margins. Integrating situational context into game evaluation adds an analytical layer that pure statistical comparison does not capture.
A team that played a night game the previous evening and then has a day game the following afternoon has had less rest, less preparation time, and often less sleep. Research across multiple seasons has shown that teams in this spot see a modest but persistent decline in both offensive production and starting pitcher effectiveness. The effect is strongest when the day game features a different starting pitcher than the previous night's game, meaning the coaching staff has less time for advance scouting preparation as well.
Travel effects are measurable in baseball, particularly when teams cross multiple time zones. A team flying from the West Coast to the East Coast for a series opener loses three hours of perceived time and often arrives late the previous night. Research has documented a small but consistent decline in performance for visiting teams in series openers following cross-country flights, particularly in first-pitch timing windows (early afternoon East Coast starts are hardest on West Coast teams). The effect diminishes after the first game of a series as bodies adjust.
Teams generally align their rotation to deploy their best starter in series openers, especially in divisional matchups. This means series openers tend to feature stronger pitching matchups than games two and three. Conversely, series finales often feature the back end of the rotation and may also involve lineup rest days as managers look ahead to the next series. Understanding where you are in a series matters for both the pitching and lineup context.
Teams within the same division face each other 13 times per season in the current format. By mid-season, divisional opponents have deep familiarity with each other's pitching staffs, lineup tendencies, and bullpen deployment patterns. This familiarity tends to compress outcomes, making divisional games more tightly contested on average. Non-divisional and interleague matchups, where familiarity is lower, often produce wider margins.
For an in-depth look at all major situational angles with historical data, see the Situational Betting Guide.
The standard MLB run line is set at 1.5 runs. The favorite is typically listed at -1.5 with plus-money odds, meaning they must win by two or more runs. The underdog is listed at +1.5 with minus-money odds, meaning they can lose by one run and still cover. Deciding between the moneyline and the run line is one of the most important structural decisions in baseball analysis, and the correct choice depends on the probability distribution of scoring margins for a given game.
Approximately 30% of all MLB games are decided by exactly one run. This means that roughly 70% of games are decided by two or more runs. When a strong favorite is expected to win by a wide margin, such as when an elite starter faces a weak lineup in a hitter-friendly park, the run line at -1.5 may offer a better return than the moneyline because the probability of a multi-run victory is high relative to the odds being offered. The moneyline in these situations may be heavily juiced (say, -220), while the run line may be offered at plus-money (say, +110). If the true probability of the favorite winning by 2+ runs exceeds the implied probability of the run line odds, the run line is the higher expected value selection.
In games where the matchup is tightly contested, where the pitching is dominant on both sides, or where the park suppresses offense, one-run games become more likely. In these scenarios, taking the underdog moneyline rather than laying -1.5 on the favorite captures the full value of the win probability without requiring a margin. Conversely, taking the favorite on the moneyline rather than at -1.5 avoids losing a selection that the correct team won but only by one run.
Over the last decade of MLB data, roughly 30% of games are decided by one run, 20% by two runs, 15% by three runs, and the remaining 35% by four or more runs. These distributions shift based on run environment: high-scoring games are more likely to produce multi-run margins, while low-scoring games cluster toward one-run outcomes. Pitching-dominated matchups at pitcher-friendly parks skew heavily toward one-run margins, which favors moneyline evaluation over run line evaluation.
For a detailed decision framework with probability tables and scenario analysis, see the Run Line vs Moneyline page.
Before selecting a run line, ask two questions: (1) What is the approximate probability that this game is decided by two or more runs? (2) Do the offered odds on the run line reflect that probability, or do they exceed it? If the odds exceed the true probability, the moneyline or the other side of the run line is the better structural choice.
The betting market is not a prediction mechanism. It is a price-setting mechanism designed to balance liability across outcomes. Understanding how lines are set, how they move, and what information the market contains is a distinct analytical discipline, separate from evaluating the game itself. Market literacy is as important as statistical literacy in the context of baseball analysis.
Prices originate from a small number of market-making sportsbooks that employ quantitative models and professional oddsmakers. These initial prices reflect the market maker's probability estimate for each outcome, plus a built-in margin (the "vig" or "juice"). Once prices are posted, the broader market reacts. Initial activity from sharp accounts, those with documented track records of predictive accuracy, provides the first round of price correction. By the time a price reaches the general public, it has already been adjusted by informed market participants.
Closing Line Value (CLV) is the most widely accepted benchmark for evaluating the quality of an analytical process over time. The closing line represents the market's final, most efficient price, incorporating all available information. If an evaluation consistently identifies a position before the market moves in that direction, meaning the line at the time of the evaluation is more favorable than where it eventually closes, that process is demonstrating predictive value. CLV is a process measure, not a results measure, and it is considered more reliable than win rate alone for assessing long-term analytical accuracy.
Public activity refers to recreational money, which tends to follow popular teams, recent results, and narrative. Sharp activity refers to money from professional or semi-professional operations, which tends to follow statistical models and market inefficiency identification. When public money and sharp money disagree, line movement can provide information about where the market's informed participants see value. This does not mean following sharp money blindly, but it does mean understanding which side of the market is being driven by informed analysis versus casual sentiment.
For a deeper exploration of market dynamics, see the Public vs Sharp Money guide and the Reverse Line Movement Guide.
First five innings (F5) lines isolate the portion of the game most directly influenced by the starting pitchers. By evaluating only the first five innings, F5 analysis removes the bullpen variable entirely, which can be advantageous in specific situations where the bullpen introduces significant uncertainty into the game evaluation.
In most games, starting pitchers are responsible for the first five innings. F5 results therefore reflect the matchup between the two starters, the lineups, the park, and the weather, without contamination from bullpen quality, deployment decisions, or fatigue. This is valuable when one or both teams have bullpens that are difficult to evaluate, whether due to roster turnover, recent heavy workload, or inconsistent performance. F5 evaluation provides a cleaner signal by focusing on the most knowable part of the game.
Bullpen performance is inherently higher-variance than starting pitcher performance because relief appearances are shorter and therefore noisier. A single bad inning from a reliever can completely change a game's complexion. Over a large sample, bullpen quality stabilizes, but on any given night, the variance is substantial. F5 evaluation sidesteps this variance entirely. When you have high conviction in a starting pitcher evaluation but low conviction in the corresponding bullpen evaluation, the F5 line allows you to express that conviction without taking on the bullpen risk.
F5 evaluation is most useful in three scenarios. First, when one team has a dominant starter but a weak or depleted bullpen. The full-game line may be compressed because the market recognizes the bullpen risk, but the F5 line reflects the starter's dominance more directly. Second, when both teams have unreliable bullpens, making the full-game outcome highly volatile. F5 provides a more stable evaluation target. Third, in high-total games where both starters are expected to exit early, the F5 line captures the starting-pitcher window before the game potentially devolves into a bullpen contest with unpredictable scoring patterns.
For a complete F5 evaluation methodology, see the First Five Innings Guide.
No analytical framework produces value if it is not paired with disciplined bankroll management. Baseball is a sport with extreme day-to-day variance. Even the best evaluation processes will produce extended losing stretches. The difference between a process that survives those stretches and one that does not is entirely a function of how capital is allocated.
The "unit" system is a standardized approach to position sizing. One unit represents a fixed percentage of the total bankroll, typically between 1% and 3%. A bankroll of $5,000 with a 2% unit size means each standard position is $100. Multi-unit positions, used when conviction is highest, should still be capped at a maximum of 3 to 5 units per individual game. The purpose of fixed unit sizing is to prevent emotional escalation, where losses lead to larger positions in an attempt to recover, which is the most common path to bankroll depletion.
Baseball has more inherent randomness than most major sports. The best team in any given season wins roughly 60% of its games. Even significant edges, when accurately identified, produce modest win rates over large samples. This means that short-term results, whether positive or negative, carry very little information about the quality of the underlying process. A 10-game losing streak does not mean the process is broken, and a 10-game winning streak does not mean the process is infallible. The minimum sample for drawing any meaningful conclusion about process quality is generally 200 to 300 evaluated games, representing roughly two full months of daily baseball.
In any probabilistic domain, short-term results are dominated by variance. A correct evaluation of a 58% probability event will lose 42% of the time. Over 10 trials, it is entirely possible, even common, for a correct evaluation to produce more losses than wins. Over 500 trials, the expected win rate will converge toward the true probability. This is why process discipline, meaning adherence to a consistent, analytically grounded evaluation framework regardless of recent results, is the defining characteristic of long-term success in baseball analysis. Changing the process after a small number of negative outcomes is the analytical equivalent of overfitting a model to noise.
Maintaining detailed records of every evaluation, including the reasoning, the metrics consulted, the line at the time of evaluation, and the final result, is essential for process review. Over time, these records reveal patterns: which analytical pillars are contributing the most value, where blind spots exist, and whether the process is genuinely identifying value or benefiting from variance. Without tracking, there is no mechanism for self-correction, and the process stagnates.
For a dedicated guide on bankroll structures and risk management frameworks, see the Bankroll Management Guide.
The following pages expand on every topic covered in this guide. Each is a standalone educational resource designed for reference use.