MLB Indicator Framework
A governed system for identifying which baseball-specific conditions may influence a BrownBagBets decision, how they relate to one another, and what the current evidence is actually teaching us about their use.
Baseball analysis begins with structure, not statistics.
BrownBagBets organizes MLB information into five indicator families so that game conditions can be evaluated without confusing volume, overlap, or narrative with genuine decision value.
The framework does not assume that every indicator matters in every game. A pitcher-outs wager, first-five side, full-game total, and batter prop may require completely different evidence even when they involve the same teams.
The purpose of the register is to identify the relevant questions first, then determine whether qualified and independent evidence supports an actionable interpretation.
What we are learning from real BrownBagBets decisions.
The MLB Indicator Framework is now being observed through actual published BrownBagBets decisions. The goal is not to crown winning signals. It is to see how registered indicators behave when they enter real theses, real prices, real uncertainty, and real outcomes.
Each qualifying indicator use is recorded as observational research. Wins and losses remain visible. Closing line value is reviewed where coverage exists. Sample size and evidence quality determine how much confidence the record deserves.
Evidence is accumulating, but most parent indicators are still early.
The current August register contains 107 supporting indicator uses across the twelve MLB parent categories. That does not mean 107 independent signals and it does not represent betting performance. A single wager may contain several indicators, and some may answer the same underlying question.
Starting Pitcher Current Condition has the largest current sample. Most other indicators remain too early to rank, while two parent categories have no qualified sample yet.
Observed indicator uses across the current August research register.
Starting Pitcher Current Condition supporting uses.
The approved MLB framework remains unchanged.
Defense and Catcher Impact; Ballpark and Weather Conditions.
Observed so far. Not ranked. Not promoted as proven edges.
These cards summarize how each parent indicator has appeared in the current August dataset. The numbers describe recorded supporting uses, not isolated predictive power.
Starting Pitcher Current Condition
DevelopingIt is the most frequently used MLB indicator in the current dataset and appears to be a legitimate core evidence input.
Limitation: It repeatedly fails inside weak or incomplete broader theses and should not be treated as a standalone predictive signal.
Expected Game Shape
Too EarlyOne of the more interesting early observations, especially because the parent category contains several different game-shape mechanisms.
Limitation: Ten uses are far too few to present 80% as a proven edge. The research question is which sub-mechanisms are actually transferring.
Bullpen Availability and Quality
Needs Cleaner EvidenceThe numerical result is interesting, but the evidence quality is not clean enough to support a responsible performance claim.
Limitation: Several bullpen observations were not recorded as fully pregame confirmed. Evidence quality takes priority over an attractive record.
Lineup Matchup Advantage
Too EarlyThis remains useful contextual evidence, but it is not separating meaningfully on outcomes in the current sample.
Limitation: Persuasive split and matchup statistics have still supported losing wagers. A good matchup statistic does not automatically create a good wager.
Offensive Approach Matchup
Too EarlyHigh research interest because the mechanism is often behavioral, including chase, walk, strikeout, and workload interaction.
Limitation: Five supporting uses are a very small sample. Continue collecting before assigning meaningful weight.
Market Positioning and Price Behavior
Too EarlyThis is strategically important because it gives BrownBagBets another way to evaluate decision quality besides the final score.
Limitation: The current sample does not establish predictive power. Price behavior is a research lens, not a guarantee that the wager wins.
Starting Pitcher Matchup Fit
Too EarlyContinue collecting. The current record is observational only.
Limitation: Six uses are too few to rank, dismiss, or promote the indicator.
Starting Pitcher Workload and Removal Risk
Too EarlyContinue collecting.
Limitation: There is too little evidence to draw a useful conclusion.
Confirmed Lineup and Player Availability
Too EarlyInteresting but unproven. Continue collecting.
Limitation: Eight supporting uses are not enough to infer a durable edge.
Schedule, Travel, and Rest
Too EarlyContinue collecting without treating schedule narratives as automatic signals.
Limitation: The current sample is too small to separate useful fatigue effects from narrative noise.
Defense and Catcher Impact
No Qualified SampleNo qualified August sample yet.
Limitation: No performance interpretation should be made without recorded qualifying uses.
Ballpark and Weather Conditions
No Qualified SampleNo qualified August sample yet.
Limitation: The framework recognizes the category, but the current register does not support a public performance claim.
Observed records are not the same thing as validated predictive edges.
A small winning sample can be interesting and still deserve very little confidence. A losing sample can be informative without proving the indicator is useless.
- Indicator support is not the same as isolated indicator causality.
- Several indicators can appear in the same wager, so their records are not independent performance records.
- CLV coverage is incomplete for some categories.
- Some bullpen observations require cleaner pregame confirmation before public conclusions deserve more weight.
- No current parent-indicator sample should be described as statistically significant or proven.
Use the current sample to ask better questions, not to manufacture certainty.
The framework separates where information comes from before asking what it means.
Starter condition, fit, and expected workload
Current condition, arsenal interaction, command, contact profile, pitch efficiency, and removal risk.
Who is playing and how the lineup attacks
Confirmed availability, batting order, matchup advantage, offensive approach, depth, and run-creation pressure.
What happens after contact and after the starter
Bullpen condition, defensive alignment, catching, relief hierarchy, and manager usage.
The conditions surrounding the matchup
Park, weather, schedule, travel, umpire, roster movement, and expected game shape.
What the market expects and what remains available
Price behavior, probability comparison, news absorption, cross-market consistency, and historical fit.
The public framework begins with twelve parent indicators.
Each parent indicator can later connect to a deeper page containing its definition, mechanism, market applicability, evidence requirements, overlap risks, and limitations.
Starting Pitcher Current Condition
Evaluates the pitcher’s present physical, mechanical, workload, velocity, command, and contact-quality condition without relying on surface results alone.
30 August supporting uses · 19–11 · developing evidence
Future indicator pageStarting Pitcher Matchup Fit
Tests whether a pitcher’s arsenal, movement, velocity bands, contact profile, and platoon strengths fit the particular opposing lineup.
6 supporting uses · 2–4 · too early to rank
Future indicator pageStarting Pitcher Workload and Removal Risk
Estimates how likely the starter is to remain in the game based on pitch counts, health, buildup, manager behavior, bullpen conditions, and opponent approach.
4 supporting uses · 2–2 · requires more observation
Future indicator pageConfirmed Lineup and Player Availability
Uses the official lineup, batting order, catcher assignment, rested regulars, injury limits, promotions, trades, and bench options.
8 supporting uses · 5–3 · interesting but unproven
Future indicator pageLineup Matchup Advantage
Evaluates how the confirmed lineup’s handedness, pitch-type performance, power, contact zones, depth, and role distribution interact with the opposing staff.
17 supporting uses · 9–8 · contextual evidence, not a demonstrated standalone signal
Future indicator pageOffensive Approach Matchup
Examines how chase, swing, contact, walk, strikeout, count management, batted-ball tendencies, and baserunning pressure affect the selected market.
5 supporting uses · high interest, very low confidence
Future indicator pageBullpen Availability and Quality
Combines recent usage, role availability, relief quality, leverage hierarchy, option moves, and expected manager access to the bullpen.
11 supporting uses · 9–2 · pregame confirmation standards need tightening
Future indicator pageDefense and Catcher Impact
Evaluates positional defense, range, conversion quality, framing, throwing, blocking, game calling, and fit with the expected contact profile.
0 qualified August supporting uses
Future indicator pageBallpark and Weather Conditions
Examines park dimensions, roof status, temperature, wind, humidity, air density, precipitation, and their interaction with the expected batted-ball profile.
0 qualified August supporting uses
Future indicator pageSchedule, Travel, and Rest
Evaluates travel, turnaround, series placement, day-night sequencing, doubleheaders, rest disadvantages, and roster-management pressure.
4 supporting uses · 2–2 · continue observing
Future indicator pageExpected Game Shape
Synthesizes the likely development of the game: pitcher-controlled, contact-heavy, bullpen-driven, high-variance, one-sided, or leverage-dependent.
10 supporting uses · 8–2 · sub-mechanisms need separation
Future indicator pageMarket Positioning and Price Behavior
Evaluates the current number, market movement, public participation, news absorption, cross-market behavior, and the difference between price and estimated probability.
12 supporting uses · 11 positive CLV observations among covered decisions
Future indicator pageMore indicators does not mean more confidence.
Indicator count is not evidence quality. Four indicators are not automatically stronger than three. The correct question is whether the indicators answer different decision questions.
Is the starter actually performing well now?
Starting Pitcher Current Condition can answer a present-state question.
Does the opponent create a favorable workload environment?
Offensive Approach Matchup can answer a different behavioral question.
Does the current number still offer participation value?
Market Positioning and Price Behavior can answer a separate market question.
What could invalidate the interpretation?
Uncertainty recognition can expose what the favorable evidence may be missing.
Four statistics describing the same recent trend may simply be one underlying signal repeated four times. A strong stack contains independent mechanisms, not decorative confirmation.
Where the signal did not transfer.
Losing decisions remain part of the research record. The objective is to identify whether the failure came from the evidence, the interpretation, the price, the uncertainty, or normal outcome variance.
A large stack still contained missing questions.
Several supporting indicators existed, including Starting Pitcher Current Condition, Starting Pitcher Matchup Fit, and Lineup Matchup Advantage. The wager lost.
Later Decision Card review exposed that the Nola evidence may not have been complete enough and that the price itself was not clearly established as value.
A large indicator stack can still contain missing evidence or weak value justification.
Positive CLV did not make the thesis automatically correct.
The wager lost despite positive closing line value. A contradictory indicator showed that Painter had allowed only three earned runs in ten innings across his two starts since being recalled.
A loss does not automatically mean the thesis was poor, and positive CLV does not automatically mean the thesis was correct. Preserve both the evidence and the contradiction.
These are not additional MLB signals. They are controls applied to every signal.
Indicator Relevance
Does the indicator actually apply to this market, player, team, and decision?
Evidence Quality
Is the information reliable, current, properly sourced, and strong enough to deserve influence?
Indicator Independence
Does the indicator add a distinct mechanism, or does it repeat information already counted?
Sample Reliability
Is the sample relevant, stable, explainable, and appropriately weighted for the claim being made?
Value Convergence
Do context-derived and price-derived value point toward the same participation decision?
Uncertainty Recognition
What remains unknown, unstable, conflicting, or capable of invalidating the interpretation?
Sample size determines how much confidence an observed record deserves.
BrownBagBets does not promote an MLB indicator as meaningful because it won several wagers. An indicator begins as an approved research category. Each qualifying use is recorded. Supporting and contradictory evidence are preserved. Outcomes and closing prices are reviewed where data is available.
The sample is insufficient for meaningful public confidence. Continue collecting.
The sample is large enough to support structured observation, but not a proven edge claim.
A future maturity state for evidence with enough quality, consistency, and sample depth to justify deeper formal review.
These public labels describe research maturity only. They do not imply statistical significance unless it has actually been established.
The indicator register is a prerequisite to stacking.
The system first identifies the relevant registered indicators, then verifies the evidence, removes overlap, and only then combines the remaining inputs.
Select applicable indicators
The sport and market determine which parent and child indicators may enter the decision.
Apply evidence standards
Each input must earn influence through reliable and decision-relevant evidence.
Remove duplication
Shared mechanisms and derivative statistics are grouped so they are not counted as separate confirmation.
Build the evidence structure
Independent indicators combine to shape expected game form, value, confidence, uncertainty, and price discipline.
The framework should become more precise as the evidence gets cleaner.
The next stage of research is not simply collecting more wins and losses. It is improving mechanism definition, confirmation standards, contradiction handling, and confidence calibration.
Whether Expected Game Shape breaks into more useful sub-mechanisms instead of remaining one broad parent interpretation.
Whether Bullpen Availability remains interesting once pregame confirmation standards are tightened.
Whether Offensive Approach Matchup becomes more useful for player props than for team markets.
Whether Market Positioning and Price Behavior continues producing positive CLV across a larger and more complete sample.
Whether Starting Pitcher Current Condition performs differently when paired with genuinely independent supporting indicators.
Whether explicitly recorded contradiction improves confidence calibration and capital allocation.
The framework does not ask how many favorable facts exist. It asks which governed indicators are present, independent, relevant, and not already priced in.
A recognized MLB indicator authorizes a line of inquiry. It does not guarantee predictive power, justify a wager, or receive equal weight in every market. Operational influence must always be earned through evidence, independence, price, and review.