Monthly Return Reliability

Monthly Return Reliability measures how reliably the selected trades reproduced a dependable monthly R result when opportunities were available. It is designed for traders whose monthly opportunity count changes and whose results may include occasional runners.

The score is available with Pro on Risk and contributes directly to the Reliability dimension of Trading Mastery.

Why simpler approaches are insufficient

Average monthly R can be lifted by a few exceptional months, making ordinary profitable months look weak. Average R per trade ignores whether enough opportunities occurred to produce meaningful monthly output. Pure smoothness can reward consistently losing months.

Monthly Return Reliability uses a hybrid:

1. Find a cautious monthly R floor from completed active months.
2. Estimate each month's opportunity from trade count and observed R expectancy.
3. Lower the target when few opportunities occurred.
4. Measure downside-only shortfall beneath that target.
5. Scale the miss by normal variation in the selected trade outcomes.
6. Give every eligible month equal final weight.

Eligible months and observations

Reliability uses complete active calendar months in the user's local timezone.

• A month is active when selected executed trades closed in it.
• Inactive months are neutral and excluded.
• Partial first and last boundary months are excluded.
• The current month is always excluded, including on its final local day.
• An active month with no observed R is missing data, not inactive.
• Monthly Net R is the sum of observed realized R.

This differs from Sharpe and Sortino, which insert inactive complete months as 0% observations to preserve elapsed calendar time.

Dependable Monthly Return

Dependable Monthly Return is the lower quartile of eligible active-month Net R:

Dependable Monthly Return = Q25(active-month Net R)

It answers: what monthly R result was met or exceeded in approximately 75% of eligible active months?

Trading Vault uses the Type-7 interpolated percentile. After sorting n monthly results:

position = (n - 1) * 0.25
lower index = floor(position)
upper index = ceil(position)
fraction = position - lower index

Q25 = lower value + fraction * (upper value - lower value)

Worked quartile example

For these 12 sorted monthly returns:

-2.99, -0.08, 0.40, 2.28, 2.29, 4.19,
 4.43,  4.67, 6.24, 7.40, 8.25, 13.99

The position is 2.75, between 0.40R and 2.28R:

Q25 = 0.40 + 0.75 * (2.28 - 0.40) = 1.81R

Interpolation prevents one exact month from abruptly controlling the floor as filters or new observations change.

Why not the mean?

Consider:

+2R, +2R, +3R, +2R, +10R, +7R, +2R, +5R, +2R, +8R

The repeatable floor is near +2R. The +10R and +8R runners should not raise the target enough to penalize normal +2R months.

A non-positive dependable floor remains visible, but it does not demonstrate reliably positive production. The final score is capped at 50 when either the dependable floor or overall R expectancy is non-positive.

Opportunity-adjusted monthly target

Overall R expectancy estimates expected result per observed opportunity:

expectancy = sum(observed trade R) / observed R trade count

Let:

F = max(0, Dependable Monthly Return)
mu = max(0, overall R expectancy)

For an active month with n trades:

opportunity expectation = n * mu
adjusted target = min(F, opportunity expectation)

If the dependable floor is +10R and expectancy is +1.5R per trade:

This prevents a low-opportunity month from being held to the full normal floor. Once enough opportunities occurred, the target stops increasing. Inactive months receive no target.

Trade expectancy is used only to estimate opportunity. A larger expectancy does not independently award more Reliability points.

Downside-only shortfall

shortfall = max(0, adjusted target - monthly Net R)

A month meeting or exceeding target has zero shortfall. Upside above target never offsets another month's miss and never creates a penalty. A traded break-even month is active and records a shortfall when its target is positive.

Average Monthly Shortfall is:

sum(monthly shortfalls) / eligible active month count

Months meeting target contribute zero. It is a downside-only average, not the signed difference between every result and target.

Normal trade-outcome variation

Different strategies have different outcome distributions. Reliability scales shortfall using the population standard deviation of observed trade R:

trade R variation = population SD(observed trade R)

trades required for floor = F / mu
effective opportunities = min(month trade count, trades required for floor)

normal monthly variation = trade R variation * sqrt(effective opportunities)

The effective opportunity count stops increasing once the dependable floor is reachable. A high-volume month therefore does not receive unlimited tolerance after its target has already been capped.

When normal variation and shortfall are both zero, standardized shortfall is zero. When shortfall is positive but valid variation is zero, the month receives the maximum shortfall classification.

Monthly score

standardized shortfall = shortfall / normal monthly variation

Trading Vault maps the result using linear interpolation:

This is deliberately one-sided. Strong positive months receive full credit without making weaker months look more reliable.

Headline Reliability score

Monthly Return Reliability =
    sum(eligible monthly scores) / eligible active month count

Every eligible month has equal final importance after opportunity count and expected variation have been accounted for. A high-volume month does not receive more final weight merely because it contains more trades.

The result is displayed from 0–100 without a percentage sign or R suffix. It is a score, not an estimated probability.

• 100 means every eligible active month met its opportunity-adjusted target.
• A high score means shortfalls were absent or modest relative to normal variation.
• A middle score reflects mixed delivery, meaningful misses, or the non-positive edge/floor cap.
• A low score reflects repeated or severe deterioration beneath adjusted targets.

Supporting evidence

Target Attainment counts months where Net R met or exceeded the adjusted target:

attainment rate = attained months / eligible active months * 100

Active-Month Participation shows how many selected months contained opportunities. It is context only and does not multiply or cap Reliability.

The monthly chart compares Net R with the adjusted target. Inactive months should be read as neutral gaps, not failed targets. Evidence may also show R expectancy, R observations and coverage, trade R variation, excluded partial months, and missing observations.

Worked scenarios

Same process, different opportunity counts

January:  2 trades,  +4R total
February: 8 trades, +16R total
March:    1 trade,   +2R total

Lower-opportunity months receive smaller targets. February's greater production does not inflate the dependable floor through a simple mean or give February extra final weight.

Stable delivery from different combinations

January:  +6R, -1R                         = +5R
February: +2R, +2R, +2R, -1R, 0R          = +5R
March:    eight mixed outcomes             = +5R

Per-trade averages differ, but monthly delivered production is dependable. Reliability can remain high.

Occasional losing months

+9R, +10R, -2R, +12R, +8R, +11R, -4R, +10R

The losing months create shortfalls. Their impact depends on frequency, severity, adjusted targets, and normal trade variation. Positive months above target do not cancel them.

No-trade versus break-even month

A month with no selected executed trades is inactive and neutral. A month with executed trades summing to 0R is active and can miss a positive adjusted target.

Evidence requirements

There is no minimum trade count for an individual month. Six to eleven active months or 30 to 49 observed trades can produce a limited result when every hard requirement is met. Below a minimum, the score is unavailable.

Missing R values are excluded rather than converted to zero. An active month with trades but no observed R makes the required history incomplete.

Relationship to other statistics

• R expectancy measures average edge per trade. Reliability uses it only to adjust opportunity.
• Return smoothness measures stability of monthly percentages and can score constant losing months highly.
• Sortino measures mean percentage return relative to downside deviation.
• Calmar compares annualized percentage growth with maximum drawdown.
• Average monthly R can be lifted by exceptional months. The dependable floor deliberately resists that effect.

Review these together on Risk. Reliability adds a specific view of repeatable active-month R production. It does not replace profitability, drawdown, or risk-adjusted return analysis.