BetTune: Filter Metrics

Structure of Metric Names

In the BetTune system, each metric has a compound name that explains its essence and calculation method:

  • Square brackets [...] at the beginning indicate the category or subject of the metric. For example:

    • [Odds] — odds related to match results or totals.
    • [Match] — metrics calculated at the match level (considering both teams).
    • [Tournament] — statistics at the tournament level.

  • Curly braces {...} clarify the calculation method or data selection:

    • {Total} — the metric is calculated on the full sample without splitting into home/away matches (all matches of the team).
    • {Home/Away Split} — statistics are split by home and away matches: for the home team, home stats are taken; for the away team — away stats.

  • Parentheses (...) indicate sample restrictions:

    • (last 5 matches) — calculated based on the last 5 matches. This helps to capture short-term trends caused by, for example, injuries to key players (in attack or defense), which directly affect the team’s ability to create or prevent chances. It also shows deviation from seasonal averages: if the value of a metric over the last 5 matches significantly differs from the seasonal value, this may signal important changes to consider when analyzing the total.
    • (last 7 matches) — same as above but for the last 7 matches.

  • Additional suffixes in metric names:

    • 0.3max — means “xG trimming”: each individual moment with xG > 0.3 is trimmed to 0.3. This reduces the impact of rare but volatile events like penalties or open-goal shots and makes xG-based metrics more stable.

Basic Match Information and Odds

  • Odds 1 – Odds for a home team win (outcome “1”). This is the bookmaker’s decimal odds that the home team will win the match.
  • Odds X – Odds for a draw (outcome “X”). Decimal odds for the match ending in a draw.
  • Odds 2 – Odds for an away team win (outcome “2”). Decimal odds for the away team to win.
  • Odds O2.50 – Odds for over 2.5 total goals (i.e., 3 or more goals in the match).
  • Odds U2.50 – Odds for under 2.5 total goals (i.e., 0, 1, or 2 goals combined).
  • Date – The match date. This parameter is not a metric in the traditional sense, but important reference information for users.
  • Tournament – The name of the tournament in which the match takes place (e.g., English Premier League, Italian Serie A, etc.). This is contextual information: different tournaments tend to have different characteristics (goal averages, draw percentages, etc.).

Tournament Metrics

[Tournament] Home Win Rate

The percentage of matches that ended with a home team victory.

This metric is calculated using the formula: (number of home wins / total number of matches).

The value ranges from 0 to 1 and is based on all available past seasons in the database for that tournament. It reflects the strength of the home-field advantage in the league.
For example, if Home Win Rate = 0.46, it means that in 46% of matches, the home team won — indicating a league with a strong home advantage.
A high Home Win Rate implies that betting on home teams tends to be more successful than in leagues with a lower rate. The average for this metric is ~0.43.
This metric is useful for tailoring strategies based on how much the home-field factor impacts outcomes — distinguishing between leagues where it plays a major role and those where it’s nearly irrelevant.

[Tournament] Draw Rate

The percentage of matches in the tournament that ended in a draw.

Formula: (number of draws / total number of matches)

The value ranges from 0 to 1 and is based on all available past seasons in the database for that tournament.
This metric reflects how frequently draws occur in the league.
For example, a Draw Rate of 0.3 means that 3 out of 10 matches ended in a draw — which is relatively high.
The average value for this metric is around ~0.26.
This metric is useful for adapting strategies based on how common draws are in a given tournament — whether draws play a significant role in determining outcomes or are relatively rare.
It helps refine and calibrate draw-focused strategies with greater accuracy.

[Tournament] Away Win Rate

The percentage of matches won by the away team in the tournament.

Formula: (number of away wins / total number of matches)

The value ranges from 0 to 1 and is based on data from all past seasons available in the database for this tournament.
A high Away Win Rate (e.g., 0.35 or above) indicates that away teams often won in this league — 35% of the time — suggesting that home advantage is not always decisive.
This helps understand the competitive balance in the tournament:
if the Away Win Rate is high, away teams frequently take points; if it’s low (e.g., 0.25), the league is “home-dominant,” and winning away is more difficult.
The average value for this metric is approximately ~0.31.
This metric is useful for tailoring strategies based on how much the away status affects a team’s chances — where playing away significantly lowers the chance of winning, and where it has minimal impact.

[Tournament] Total Over Rate

The percentage of matches in the tournament where 3 or more goals were scored.

Formula: (number of matches with 3+ goals / total number of matches)

The value ranges from 0 to 1 and is based on data from all past seasons available in the database for this tournament.
For example, a value of 0.55 means that in 55% of matches, at least 3 goals were scored — indicating a high-scoring league.
In contrast, a low value suggests a more defensively oriented competition.
The average value for this metric is around 0.43.
This metric is useful for tailoring strategies based on scoring tendencies — where the probability of high totals (Over 2.5 goals) is high, and where it is low.

Tournament Stage Metrics

[Match] Finished Matches (Average)

The average number of matches already played by both teams at the current stage of the tournament.
This metric reflects how many games each team has completed (on average) so far in the competition.
If both teams have played 6 matches, then [Match] Finished Matches (Average) will be 6.
If one team has played 9 and the other 6, the average will be 7.5.
Use this metric to exclude early rounds from your analysis and strategies, when team data is still unstable and models are less reliable.

[Match] Stage Matches Left (Average)

The average number of matches remaining for both teams in the current stage of the tournament.
It indicates how many games (on average) are still to be played by each team in this phase.
If both teams have played the same number of games, the value will be identical.
If one team has played more or fewer games, the average is calculated.
Use this metric to exclude final rounds from your analysis and strategies, where team behavior may differ significantly — due to increased motivation (e.g. title race, European qualification, relegation fight) or lack of motivation.

Expected Goals (xG)

xG – expected goals, a metric of chance quality. It estimates how many goals a team was expected to score based on the quality of chances.
xGA – expected goals against, i.e., how many goals a team was expected to concede.
xG90 / xGA90 – these are xG and xGA metrics normalized per 90 minutes (i.e., per match).

[Match] {Total} Home Team xG90

Expected goals per 90 minutes for the home team (average across all previous tournament matches).
This is the average xG generated by the home team per game.
For example, xG90 = 1.50 means the team creates chances worth 1.5 expected goals per match.
A high value indicates attacking strength; a low one signals weak offense.
Average value for this metric is ~1.22.

[Match] {Total} Home Team xGA90

Expected goals against per 90 minutes for the home team.
This reflects the average xG created by opponents against this team per match.
xGA90 = 0.90 means the defense allows opponents chances worth 0.9 xG per match.
Lower values suggest solid defense, higher ones indicate defensive vulnerability.
Average value: ~1.22.

[Match] {Total} Home Team xG90 + xGA90

Total xG per match involving the home team (sum of their xG and xGA per match).
For example, xG90 = 1.5 and xGA90 = 1.0 gives a total of 2.5, indicating high-scoring potential.
This shows whether games involving the team are open and attacking or closed and defensive.
Average value: ~2.44.

[Match] {Total} Away Team xG90

Expected goals per 90 minutes for the away team (average across all previous tournament matches).
It shows how productive the team is in attack.
For example, xG90 = 1.20 means the team creates chances worth 1.2 goals per game.
A high value indicates an attack that generates good chances even in away conditions.
Average value: ~1.22.

[Match] {Total} Away Team xGA90

Expected goals against per 90 minutes for the away team.
Reflects the average xG created by opponents against them per match.
xGA90 = 1.10 indicates the team allows chances worth 1.1 goals.
Lower values signal strong defensive structure.
Average value: ~1.22.

[Match] {Total} Away Team xG90 + xGA90

Total xG in matches involving the away team (sum of their xG and xGA across all previous matches).
For example, xG90 = 1.4 and xGA90 = 1.3 results in 2.7, suggesting high chance volume in away games.
High values indicate open games; low values suggest tighter contests.
Average: ~2.44.

[Match] {Total} Home(xG90 + xGA90) + Away(xG90 + xGA90)

This metric sums up:

  • the home team’s total xG90 + xGA90 based on all its matches
  • and the away team’s total xG90 + xGA90 based on all its matches.
    It gives the combined average xG activity of both teams in previous games.
    For example, if the combined value is 5.5 or higher, the match may be explosive with lots of chances.
    A value around 4.5 or lower suggests a more defensive, pragmatic game.
    A highly useful indicator when analyzing total goals markets.
    Average value: ~4.9.

How to Use

By comparing the metrics of the home and away teams, the user can understand each team’s style.
For example, if both teams have a high sum of xG90 + xGA90 (e.g., 2.5–3.0), the upcoming match is likely to feature many chances and goals.
If both teams have low values (e.g., ~1.8), the game is likely to be defensive.
It’s also important to compare attack vs defense:
if the home team’s xG90 and the away team’s xGA90 are both significantly above average,
the home side is likely to create more chances.

Expected Total xG Metrics

Expected Total metrics estimate the expected total number of goals in a match based on a comprehensive analysis of both teams’ statistics. The calculation includes average expected goals (xG90) and expected goals against (xGA90), as well as attacking and defensive strength ratings. These ratings are unitless values — they reflect relative team strength, but not the overall goal intensity of the league. That’s why they are multiplied by a scalar — the average xG90 for the season, calculated separately for each tournament. This approach allows for accurately relating team strength to the league context and provides a more objective estimate of the expected total than a simple xG sum.

[Match] {Total} Expected Total xG

The expected total number of goals (based on xG). This is the final forecast from the BetTune model: how many goals are expected on average in the match, based on both teams’ xG statistics. It uses each team’s xG90 and xGA90, attacking and defensive strength ratings, and the tournament’s seasonal average xG90. This metric serves as a reference for totals. Comparing it with the bookmaker’s total (e.g., 2.5) helps to spot mismatches. If [Match] {Total} Expected Total xG is significantly higher than the bookmaker’s total, it may suggest an “over” scenario — and vice versa. The average value of this metric is ~2.22.

[Match] {Total} Expected Total xG (last 7 matches)

This is the same metric as [Match] {Total} Expected Total xG, but based only on the last 7 matches for each team. It reflects the expected number of goals considering recent form and changes. This helps capture current trends — for instance, due to injuries of key attacking or defensive players, which directly impact chance creation or prevention. If this metric diverges from the seasonal version, it could signal important recent changes. The average value is ~2.22.

[Match] {Total} Expected Total xG 0.3max

This metric estimates the expected total goals in the match while capping the xG of each shot at 0.3. If a shot had an xG of 0.65, it is counted as 0.3 in this metric. All shots with xG ≤ 0.3 remain unchanged. This helps reduce the impact of rare, high-xG events like penalties or open-goal chances — which can distort xG totals due to their high variance. The 0.3max metric offers a more stable and repeatable view of attacking potential. It’s especially suitable for analyzing totals where excluding super-high moments adds reliability. Its average value is lower than the standard Expected Total xG: ~1.91 vs ~2.22, so thresholds shouldn’t be compared directly.

[Match] {Total} Expected Total xG, 0.3max (last 7 matches)

Same as [Match] {Total} Expected Total xG, 0.3max, but based on only the last 7 matches for each team. All shots with xG > 0.3 are capped at 0.3. This helps reflect current team form, lineup changes, injuries, or tactical adjustments. Average value: ~1.91.

[Match] {Home/Away Split} Expected Total xG

This metric calculates the expected total goals by taking into account the home/away context. For the home team, only home matches are used; for the away team — only away matches. Since teams often behave differently depending on the venue (home teams often attack more, away teams are more cautious), this provides a more accurate match-specific forecast rather than a season average. The average value is ~2.22.

[Match] {Home/Away Split} Expected Total xG, 0.3max

This version combines the home/away split and the 0.3 xG cap. For the home team, only home matches are considered; for the away team — only away matches. Shots with xG > 0.3 are limited to 0.3. This provides a more stable and realistic estimate of expected goals in typical match conditions for each team. It reduces variance by removing the effect of rare high-xG events. Average value: ~1.73.

How to Use

When building strategies in BetTune, it’s helpful to look at different versions of Expected Total xG — overall, last 7 matches, 0.3max, and home/away split.
This gives a better understanding of how expected total varies with form, venue, and reduces the noise from one-off super chances.
Such a layered approach makes the strategy more accurate and robust.

Expected Difference xG Metrics

This group of metrics estimates the expected goal difference between two teams based on their attacking and defensive potential. Unlike Expected Total (which shows the combined number of goals in a match), this metric answers the question: who is more likely to win and by what margin. The same core inputs are used — xG90, xGA90, attacking and defensive strength ratings, and average tournament xG values. However, instead of summing the teams’ contributions (as with Expected Total), this metric uses the difference in expected goals, normalized by contextual factors (home/away form, opponent strength, league scoring environment). These metrics help objectively assess one team’s edge over the other — especially valuable for analyzing win bets. The further the Expected Difference xG value deviates from zero, the stronger the signal of imbalance between the teams.

[Match] {Total} Expected Difference xG

Estimates the expected goal difference between the two teams. Used in models calculating win/draw/loss (1X2) probabilities. The higher the positive value of [Match] {Total} Expected Difference xG, the greater the chance of the first (home) team winning in the long run. A strongly negative value suggests the second (away) team is more likely to win. Values near zero imply evenly matched teams and a higher draw probability. Unlike Expected Total (which sums both teams’ xG), this metric focuses on the expected superiority of one team. Average value: ~0.18.

[Match] {Total} Expected Difference xG (last 7 matches)

Same as [Match] {Total} Expected Difference xG, but based only on each team’s last 7 matches. This allows the metric to reflect recent form and conditions — for example, lineup changes, tactical shifts, or injuries to key players. Average value: ~0.18.

[Match] {Total} Expected Difference xG, 0.3max

In this version, all shots with xG > 0.3 are capped at 0.3, reducing the impact of rare high-xG chances (e.g., penalties, open goals). This increases the long-term stability of predictions, which is essential for building robust models. Average value: ~0.

[Match] {Total} Expected Difference xG, 0.3max (last 7 matches)

Same as above, but calculated over the last 7 matches for both teams. This captures recent dynamics while filtering out unstable high-impact moments. Useful for short-term modeling with minimal noise. Average value: ~0.

[Match] {Home/Away Split} Expected Difference xG

This version uses only the home team’s home matches and the away team’s away matches in the calculation. It accounts for behavioral and tactical differences that often occur between home and away games. Many teams perform differently depending on the venue, so ignoring this factor can result in systematic forecasting errors. Average value: ~0.2.

[Match] {Home/Away Split} Expected Difference xG, 0.3max

Combines the home/away split with the 0.3 xG cap per shot. This offers a stable and realistic estimate of one team’s advantage in match-specific conditions. The higher the value, the greater the expected advantage for the first (home) team. Negative values favor the second (away) team. Values near zero indicate balanced teams. Average value: ~0.15.

How to Use

A positive Expected Difference xG value of +0.3 or more signals a statistically meaningful advantage for the first (home) team. A negative value indicates the second (away) team has the edge. Values close to zero suggest a balanced matchup and higher draw probability.

This metric is especially useful as an input for match outcome prediction models (1X2), as it directly reflects the structure and direction of expected team dominance. When used in combination with other variables — such as total goals, form, home/away splits, or variance — it helps improve the accuracy and robustness of probability-based models over the long run.

Historical Betting Valuation Difference Metrics

These metrics reflect how often the bookmaker’s opening line mispriced a team in recent matches. The calculations are based on indicators like CLV (Closing Line Value) and Positive CLV, which compare the opening odds to the closing odds. They also consider additional factors — such as when the market moved (early vs late in the line’s lifetime) and how that movement aligned with the actual match outcome. As a result, this metric doesn’t just measure betting efficiency against the closing line, but captures deeper patterns of consistent mispricing. This allows models to detect systematic undervaluation or overvaluation by the market and use that as an additional signal.

[Match] {Total} Historical Win Valuation Difference

This metric measures how accurate the market’s odds for winning were in past matches for both teams. It incorporates win-market CLV values, the direction and magnitude of odds movement, and whether the movement occurred early or late in the line. If the market consistently overvalued one team and undervalued the other, this creates a detectable bias captured by the metric. It is used in 1X2 outcome models to flag possible mispricing in the current match based on past market inefficiencies.

[Match] {Total} Historical Win Valuation Difference (last 5 matches)

Same as [Match] {Total} Historical Win Valuation Difference, but based only on each team’s last 5 matches. This helps models respond to short-term shifts in market perception. Especially useful when a team has recently changed tactics, lineups, or form, but the market hasn’t fully adjusted its opening lines yet.

[Match] {Total} Historical Total Under Valuation Difference

Measures how often the market mispriced the “total under” line in previous matches. Based on CLV and Positive CLV values for under bets, the nature of odds movement, and final outcomes. If the opening line regularly underestimated the chance of an “under,” this is flagged as a market error. Useful for models trying to correct for historical mispricing in low-scoring expectations.

[Match] {Total} Historical Total Under Valuation Difference (last 5 matches)

Same as [Match] {Total} Historical Total Under Valuation Difference, but calculated only on the last 5 matches of each team. Shows whether recent opening lines consistently undervalued the under, supported by CLV metrics and final outcomes. The metric also considers whether the line movement occurred early or late. Helps models respond to short-term tactical or form changes — for example, declining attacking performance or more conservative play — that the market hasn’t yet fully priced in.

[Match] {Total} Historical Total Over Valuation Difference

Measures whether the market systematically mispriced the “total over” line in previous matches. Based on over-market CLV values, movement patterns, and match outcomes. If the market frequently underestimated the likelihood of a high-scoring game, this is captured as a persistent bias. Helps models detect and correct for market inefficiencies related to overs.

[Match] {Total} Historical Total Over Valuation Difference (last 5 matches)

Same as [Match] {Total} Historical Total Over Valuation Difference, but based on the last 5 matches for each team. Evaluates whether recent opening lines consistently undervalued the chance of an “over,” using CLV metrics and actual results. It also analyzes when the market moved — early or late. Especially helpful when a team is trending more attacking, has tactical shifts, or recent high-scoring games, which the market hasn’t fully adapted to yet.

How to Use

Test whether the long-term or last-5-match versions of these metrics offer stronger predictive signals in your models. Often it’s best to use both: the long-term version reveals stable market inefficiencies, while the short-term version helps react to recent changes before the market fully adjusts.