The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Read an advanced sports statistic by asking four questions in order: what outcome or skill does it measure, what is its denominator and scale, what context or model adjustments are included, and who defines it. An index such as OPS+, a percentage such as true shooting, a probability such as expected goals, and a composite score are not interchangeable—even when they appear in the same table.
What “advanced” means in sports statistics
“Advanced” is a broad label, not a guarantee that a number is better than a basic box-score statistic. An advanced metric may normalize performance for league conditions, adjust for opportunity, combine several inputs, estimate the quality of an attempt, or calculate a probability from a statistical model. Its usefulness depends on the question you are trying to answer.
- Context-adjusted: The number is recalculated for conditions such as league scoring levels or ballpark effects.
- Opportunity-adjusted: A rate or per-possession value accounts for how many chances a player or team had.
- Composite: Several measurements are combined into one score.
- Expected or predictive: A model estimates what normally happens in comparable situations rather than reporting only what happened.
For example, Major League Baseball (MLB) describes OPS+ as league- and park-adjusted, the National Basketball Association (NBA) publishes possession-aware percentages and expected-shot measures, the National Football League (NFL) describes a composite passing score, and Opta defines football (soccer) statistics under its own data conventions. Always read the publisher’s definition before comparing numbers.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →A five-question method for reading any metric
- What question does it answer? Is it measuring total production, efficiency, contribution, shot quality, or the probability of an outcome?
- What is the unit? Identify whether the result is a count, rate, percentage, adjusted index, points-based score, or probability.
- What is the denominator? Look for at-bats, plate appearances, field-goal attempts, free-throw attempts, possessions, plays, shots, minutes, or another opportunity measure.
- What baseline is used? A league average, zero-point scale, percentage of attempts, or probability baseline changes how the number should be interpreted.
- What context and model owner are involved? Check adjustments for park, league environment, pace, opponent, game state, shot location, player tracking, and the provider’s model version or assumptions.
Then check the time window. A one-game rate can be descriptive but noisy; a full-season number usually has more opportunity behind it, while a multi-season sample can better describe a player’s underlying level. None of those horizons automatically predicts the next game.
#1 Best Overall
Know the scale before comparing numbers
| Scale | What it tells you | Typical reading mistake |
|---|---|---|
| Count | Total events, such as points, home runs, tackles, or goals. | Treating a high total as superior efficiency when opportunities differ. |
| Rate or percentage | Events relative to a stated denominator, such as attempts or possessions. | Ignoring the denominator size and small-sample volatility. |
| Adjusted index | Performance rescaled around a reference point after stated adjustments. | Reading the index as a literal percentage or probability. |
| Composite score | Several inputs combined according to a publisher’s formula. | Assuming the components or weights are universal. |
| Probability or expected value | A model’s estimate for an outcome in a situation or attempt. | Assuming the estimate is what must happen on that single attempt. |
Two percentages can also answer different questions. A shooting percentage describes makes relative to attempts; an expected shooting percentage describes the model-estimated make rate for the shots’ features. Put the label, denominator, baseline, and model beside the value whenever you save or share a statistic.
Baseball: reading OPS+ and expected wOBA
OPS+ is an adjusted hitting index
OPS+ starts with on-base plus slugging (OPS), then normalizes it for league conditions and external factors such as ballparks. MLB’s glossary gives the formula as 100 x (OBP/lgOBP + SLG/lgSLG - 1), where OBP is on-base percentage, SLG is slugging percentage, and the league terms supply the comparison environment. The complete park adjustment is part of MLB’s calculation even though the compact glossary formula displays the core relationship.
The scale is centered at 100. MLB states: “It then adjusts so a score of 100 is league average, and 150 is 50 percent better than the league average.” Thus, 100 means league-average adjusted offensive production, while 150 means 50 percent above that average on this index. A value of 80 indicates 20 percent below league average on the same scale; it is not an 80 percent success rate.
Free tools Windows power users keep installed
One-click scans. No signup required.
See the official definition at MLB’s OPS+ glossary. OPS+ is a description of adjusted past production. It is not a raw percentage, a probability of a hit, or a promise that the player will maintain the same level.
How to compare OPS+ responsibly
- Compare players within the same calculation convention and time period.
- Remember that the index already addresses league and park environment; do not add an informal park adjustment a second time.
- Check playing time separately. A part-time player can post a high OPS+ in fewer plate appearances than a durable everyday player.
- Use the index for offensive production, not for defense, baserunning, or total player value.
Expected wOBA estimates batted-ball outcomes
Expected weighted on-base average (xwOBA) is a model output associated with a batted ball. MLB explains that it assigns probabilities for a single, double, triple, and home run using comparable batted balls from the Statcast era, which began league-wide in 2015. The estimate uses the quality and characteristics of contact rather than waiting only for the observed result.
A hard line drive caught by a fielder can receive a stronger expected outcome than a softly hit single. That difference does not rewrite the box score; it separates contact quality from the result of one ball in play. Read the value together with the inputs and provider documentation at MLB’s expected wOBA glossary. Baseball Savant’s broader Statcast catalogue, including expected batting average and defensive measures, is documented at its metrics documentation.
Because xwOBA is model-based, two providers could produce different values if they use different inputs, comparison sets, or model versions. Keep the provider name attached to the number.
Basketball: eFG%, TS%, and expected field-goal percentage
Effective field-goal percentage (eFG%)
Effective field-goal percentage answers: how efficient was field-goal shooting after giving three-pointers their extra point value? A made three is worth more than a made two, so ordinary field-goal percentage treats those makes alike while eFG% accounts for the scoring value of the three.
NBA’s statistics glossary and FAQ explain eFG% and place it among the factors used to understand team performance. Read the exact definition at the NBA statistics glossary and the pace and four-factor context at the NBA statistics FAQ. eFG% covers field goals; it does not include free throws in the shooting denominator.
True shooting percentage (TS%)
True shooting percentage answers a broader efficiency question by incorporating two-point shots, three-point shots, and free throws. NBA’s published formula is Points / [2*(Field Goals Attempted + 0.44*Free Throws Attempted)]. The 0.44 is the coefficient in that glossary formula, not a law of basketball and not a coefficient that should automatically be transplanted into another provider’s metric.
Because TS% includes free-throw attempts, it can rank two players differently from eFG%. A player who draws fouls and converts at the line may have a larger advantage in TS% than in eFG%. Check the denominator and the provider’s definition before calling one player the better shooter.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchExpected field-goal percentage (expected FG%)
Expected field-goal percentage estimates whether a particular shot would go in based on comparable shot situations. NBA describes models using player-pose and tracking features, with separate approaches for jump shots and finishes around the rim. The estimate is not the result that must occur on one attempt. It is a model’s expected make rate for shots with similar features.
Rank #3
Use observed FG% beside expected FG% to ask a more precise question: did the player make more or fewer shots than the model expected, and over how many attempts? A gap over a handful of shots may be noise; a persistent gap over a large sample may invite investigation into finishing, shot creation, defense, or model limitations. The NBA’s explanation is at Intro to Expected Field Goal Percentage.
Football: interpreting a composite passing score
NFL advanced passing measures combine multiple aspects of a throw instead of reporting only completions or yards. The NFL Next Gen Stats introduction describes a passing-score model with several components, including completion percentage over expected (CPOE) and expected air EPA. In that article, CPOE contributes 11 percent of the score and expected air EPA contributes 7 percent.
Those percentages are weights in the specific NFL model described by that article. They are not general rules for every football rating, and the article is older than the current date. Metric definitions and weights can change, so consult the live NFL documentation before using the exact formula in a current-season analysis: NFL Next Gen Stats’ passing-score introduction.
How to read the components
- Completion percentage over expected: Completions are evaluated against the completion probability expected for the pass situation, rather than treated as identical throws.
- Expected air EPA: The model considers the expected value of the air-yard component of a pass within the model’s broader expected-points framework.
- Composite score: The final number reflects the model’s combination of components. It is not a pure completion percentage, yardage total, or independent probability.
When comparing quarterbacks, record the season, model version, sample of plays, and whether the comparison covers all attempts or a selected situation. A score from one game can be useful for describing that game but is too limited to establish a stable player level by itself.
Soccer: what expected goals (xG) means
Expected goals (xG) assigns a model-estimated scoring probability to a chance. A shot with an xG of 0.30 represents a model estimate of roughly a 30 percent scoring chance for opportunities with the relevant features; it does not mean the team receives 0.30 goals on that individual shot. Ten shots with a combined xG of 3.0 still can produce zero, one, or several actual goals.
Features can include location and other event details, but the precise inputs, training data, and treatment of situations depend on the provider. Opta publishes its own football-stat definitions, including expected goals, at Opta’s football statistics definitions. Do not assume an xG value from one provider is interchangeable with another provider’s xG unless their documentation supports that comparison.
Using xG without confusing chances and finishing
- Use total xG to describe the quality and volume of chances created.
- Compare goals with xG only after checking the number of shots and the time period.
- Separate chance creation from finishing: a team can create high-quality chances and finish poorly, or create modest chances and score unusually often.
- Keep penalties and other special situations consistent, because their treatment can materially change totals.
Why pace, possessions, and opportunity change basketball comparisons
A team that plays faster has more possessions in which to accumulate points, rebounds, and assists. NBA’s FAQ emphasizes pace and possession context, along with the four factors used to analyze team performance. Per-game totals therefore need a pace check before they are used to compare teams with different tempos.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitchesPer-possession rates reduce one source of distortion, but they do not remove every difference. Opponent quality, lineup role, game state, shot selection, and sample size still matter. A bench player’s rate in short stints is not automatically comparable with a starter’s rate across heavy minutes. Write down both the rate and the opportunity behind it.
A cross-sport comparison framework
| Metric example | Primary question | Scale | Key context to verify |
|---|---|---|---|
| OPS+ | How did offensive production compare with league average after stated adjustments? | Adjusted index centered at 100 | League, ballpark, season, and plate appearances |
| xwOBA | What outcome would comparable batted balls be expected to produce? | Model-estimated run-value scale | Statcast-era comparison data, contact features, and provider model |
| eFG% | How efficient was field-goal shooting after valuing threes correctly? | Percentage | Field-goal attempts, three-point share, and game or season sample |
| TS% | How efficient was scoring after including free throws? | Percentage | Field-goal attempts, free-throw attempts, and exact coefficient |
| Expected FG% | How likely were the attempted shots to go in given their features? | Model-estimated percentage | Tracking inputs, shot category, model version, and attempts |
| NFL passing score | How did several passing components combine under the NFL model? | Composite score | Component definitions, weights, season, and play sample |
| xG | How likely was a chance to become a goal under the provider’s model? | Probability per chance or summed expected goals | Provider, chance features, special situations, and sample |
The table is a map, not a universal ranking. An OPS+ of 150 cannot be placed on the same scale as a 60% TS%, and neither can be converted into an xG value. First decide which question matters, then choose a metric that actually answers it.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Expected statistics are estimates, not alternate box scores
Expected metrics are useful because outcomes contain randomness. A shot can hit the post, a well-struck ball can be caught, and a difficult attempt can go in. A model reduces some of that result noise by comparing the event with similar historical situations.
That does not make the model objective in every possible sense. The provider chooses the training data, features, outcome categories, and calibration method. Changes to any of those choices can change the value. When a number moves, check whether the athlete changed, the opportunity changed, or the model definition changed.
Expected values also work best with enough attempts. One expected-goal estimate on one shot is a probability statement, not evidence that a player “should have scored.” Aggregated values over many attempts are more informative, but they remain estimates rather than certainties.
Best Value
- Russell Westbrook's triple-doubles: Assessing their true worth
- Postseason success: Unraveling the mystery of player performance
- Team building strategies: Draft and free agency perspectives
- Defensive prowess: The challenge of accurate measurement
- The myth of the 'quick two' shot
Common mistakes and how to correct them
Calling every percentage a success rate
A percentage can be efficiency, frequency, or a model estimate. TS% is scoring efficiency with a stated denominator; expected FG% is a model-estimated make rate. Read the label and formula before assigning a plain-English meaning.
Ignoring opportunity
Counts rise with playing time, possessions, shots, and plays. Pair totals with attempts, minutes, or possessions, and avoid treating a small-sample rate as settled ability.
Mixing providers
Two xG or expected-shooting values can differ because their models use different features and historical data. Keep values from one provider together unless a documented equivalence exists.
Comparing unlike baselines
OPS+ uses 100 as league average. A percentage is bounded differently, and a composite score may have no intuitive zero or average. Convert neither mentally nor mathematically without the publisher’s stated scale.
Assuming adjustment means perfection
Park, pace, opponent, and tracking adjustments address specific sources of context. They do not remove every influence, and they do not turn a descriptive metric into a guaranteed forecast.
Using an old formula as if it were current
Composite models can be revised. The NFL passing-score introduction’s component weights are tied to the model described in that article; verify the current page before publishing a time-sensitive comparison.
A practical workflow for reading a statistic on a website
- Copy the exact label. Record capitalization, plus signs, “expected” wording, and whether the page says rate, percentage, index, or score.
- Open the provider glossary. Find the formula, denominator, adjustment notes, and any model-version statement.
- Write a one-sentence translation. For example: “This is an adjusted offensive index centered at 100,” or “This is a model-estimated scoring probability per chance.”
- Record opportunity and sample. Note attempts, possessions, shots, plays, minutes, plate appearances, and the date range.
- Check the comparison group. Identify league, season, competition, opponent, park, and whether the value is team- or player-level.
- Compare like with like. Use the same provider, definition, season scope, and situation filters where possible.
- State uncertainty honestly. Mention small samples, model dependence, or outdated documentation when those facts affect the conclusion.
Optional further reading
For a broader mathematical treatment spanning baseball, football, basketball, and other sports, see the sample of Analytic Methods in Sports: Using Mathematics and Statistics to Understand Data from Baseball, Football, Basketball, and Other Sports at the publisher’s preview PDF. A baseball-focused sample guide covering OPS+, WAR, and defensive metrics is available from Creighton University at A Fan’s Guide to Baseball Analytics. Check the current edition and suitability for your level before relying on either as a course text.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
