Diamond Signals: Do They Actually Outperform?

By iPresage Research · 10 min read · 2025-02-10

A 7-year audit of 312,847 diamond signals reveals a 54.1% win rate vs. 47.3% baseline, with exceptional diamonds hitting 64.2%. Full performance breakdown.

Every options scanner claims to find the best trades. At iPresage, we make a specific claim: our diamond signals -- the highest-conviction subset identified by the scanner -- outperform the broader signal universe. But claims require evidence. In this report, we present a comprehensive performance audit of diamond signals using 7 years of data, 159 million options records, and rigorous statistical methodology.

This is not a marketing exercise. We present both the strengths and limitations of diamond signals, including the conditions under which they underperform and the practical constraints that affect real-world execution.

WHAT IS A DIAMOND SIGNAL

A diamond signal is iPresage's top-tier classification for options flow that meets a set of stringent criteria. The specific criteria involve the confluence of multiple factors: unusual volume relative to open interest, favorable pricing relative to the expected value model, strong institutional flow characteristics (block trades, sweep activity, and size at the offer), and alignment with the prevailing regime and sector pulse.

Not every unusual options activity print qualifies. Of the approximately 14.2 million unique option flow events in our 2018-2024 database, only 312,847 were classified as diamond signals -- a selectivity rate of 2.2%. This selectivity is by design: the value of a signal classification is inversely proportional to its frequency.

Diamond signals are further categorized by direction (bullish or bearish), confidence tier (standard, high, and exceptional), and the EV score assigned by the expected value model. The EV score represents the scanner's estimate of the signal's edge over random, expressed as a percentage of the premium paid.

THE BASELINE: ALL SIGNALS VERSUS DIAMOND SIGNALS

We begin with the broadest comparison. Across all 14.2 million flow events in the database, the average win rate -- defined as the option closing at or above the entry price at any point within 5 trading days -- was 47.3%. This is below 50%, which makes sense: options are a negative-sum game due to theta decay and the bid-ask spread.

The 312,847 diamond signals, by contrast, posted an average win rate of 54.1% over the same 5-day window. This 6.8 percentage point improvement over the baseline is statistically significant (p < 0.001) and economically meaningful. A 54.1% win rate in options, where the average payoff on winners exceeds the average loss on losers (due to gamma), translates to a positive expected value.

The average EV score on diamond signals was +2.1, meaning the scanner estimated a 2.1% edge over the theoretical fair value of the option at the time of signal generation. For context, after accounting for typical bid-ask spreads of 3-5% on the options we track, a +2.1% EV means that approximately half of diamond signals have a positive expected value after execution costs, while the other half may not. This is an honest assessment: diamond signals provide an edge, but it is not so large that execution quality becomes irrelevant.

PERFORMANCE BY DIRECTION

Bullish diamond signals outnumbered bearish ones by a ratio of approximately 1.7 to 1 (197,384 bullish vs. 115,463 bearish), reflecting the general upward bias of equity markets and the natural tendency of options flow to skew toward calls.

Bullish diamond signals posted a win rate of 55.3% with an average EV of +2.4. Bearish diamond signals posted a win rate of 52.1% with an average EV of +1.6. Bullish signals outperformed on both metrics, but the bearish signals still showed meaningful positive edge over the baseline.

The performance gap between bullish and bearish signals is partly structural -- the market goes up more than it goes down over long periods -- but it persists even after controlling for the prevailing regime. In SURGING regimes, bullish diamonds won 61.7% versus 48.3% for bearish diamonds. In DRAINING regimes, the gap narrowed dramatically: 44.7% bullish versus 58.6% bearish. In NEUTRAL regimes, it was 55.1% bullish versus 52.8% bearish.

This regime-dependent directional asymmetry is one of the most important findings in this audit. Diamond signals are not equally reliable in both directions at all times. The regime context determines which direction has the edge.

PERFORMANCE BY CONFIDENCE TIER

Diamond signals are classified into three confidence tiers based on the internal scoring algorithm:

Standard diamonds (218,993 signals, 70.0% of all diamonds) posted a 52.4% win rate and +1.3 average EV. These form the broad base of the diamond signal population and provide modest but positive edge.

High-confidence diamonds (75,054 signals, 24.0%) posted a 57.8% win rate and +3.6 average EV. The improvement from standard to high-confidence is substantial: 5.4 percentage points of win rate and nearly triple the EV.

Exceptional diamonds (18,800 signals, 6.0%) posted a 64.2% win rate and +7.8 average EV. These are rare -- averaging about 7.4 per trading day across all tracked names -- but their performance is striking. A 64.2% win rate in options, combined with the inherent leverage of the instrument, represents a significant edge.

The monotonic improvement across confidence tiers is reassuring from a methodological standpoint. It means the internal scoring algorithm is capturing real information about signal quality, not just noise.

PERFORMANCE BY UNDERLYING

Diamond signal performance varies substantially by underlying asset. We ranked the 39 stocks most actively tracked by iPresage by their diamond signal win rate over the full 7-year period.

The top five by win rate were: AAPL (57.8%, 18,412 signals), MSFT (57.2%, 16,891 signals), JPM (56.4%, 9,847 signals), SPY (55.9%, 31,206 signals), and KO (55.6%, 5,123 signals). These names combine high liquidity, tight bid-ask spreads, and institutional participation -- all factors that favor signal quality.

The bottom five were: HYG (49.8%, 2,134 signals), LQD (50.1%, 1,847 signals), GLD (50.4%, 4,912 signals), T (50.9%, 3,621 signals), and D (51.2%, 2,847 signals). The fixed-income ETFs (HYG, LQD) and gold (GLD) underperform, likely because the flow patterns that drive diamond signals are calibrated primarily for equity dynamics. Low-volatility utilities (T, D) also show weaker performance, possibly due to lower institutional options activity.

This dispersion is important: a diamond signal on AAPL carries more historical edge than a diamond signal on HYG. Users should weight their attention accordingly.

TEMPORAL STABILITY

A critical question for any signal system is whether performance is stable over time or driven by a few extraordinary periods. We calculated the annual win rate for diamond signals from 2018 through 2024:

2018: 53.4% (a moderate year with the Q4 selloff) 2019: 56.2% (a strong trending year with low volatility) 2020: 52.7% (COVID crash and recovery -- extreme conditions) 2021: 55.8% (strong bull market with elevated retail participation) 2022: 51.9% (the bear market year -- the weakest in the sample) 2023: 54.6% (recovery year with AI-driven sector rotation) 2024: 54.8% (rate-cutting cycle with moderate volatility)

The win rate has remained positive in every year, with a range of 51.9% to 56.2%. The worst year (2022) still showed positive edge over the baseline, though the margin was thinner. The consistency across bull markets, bear markets, and the extreme volatility of 2020 is encouraging and suggests the signal captures a structural feature of options markets rather than a regime-specific anomaly.

EXECUTION REALITY: SLIPPAGE AND FILL RATES

Raw signal performance assumes execution at the midpoint of the bid-ask spread at the time the signal is generated. In practice, execution involves slippage. We estimated the impact of realistic execution assumptions on diamond signal performance.

Assuming execution at the offer (worst case for buyers), the diamond signal win rate drops from 54.1% to 51.8% and the average EV drops from +2.1 to +0.3. The edge survives but is substantially reduced. At the midpoint, performance is as reported. Using limit orders at the midpoint or better, which fill on approximately 72% of signals based on our analysis of intraday price action, the win rate on filled orders rises to 55.6% with +2.9 EV -- better than the reported figures because unfilled orders tend to be the ones that moved away quickly, representing the weaker setups.

This execution analysis highlights a key practical consideration: diamond signals are not a mechanical system where execution quality is irrelevant. Patient execution using limit orders at or near the midpoint materially improves real-world outcomes.

COMPARISON TO ALTERNATIVES

We compared diamond signal performance to two common alternatives: trading options based on simple moving average crossovers, and trading the highest-volume options each day.

Moving average crossover signals (50-day crossing 200-day, applied to the same universe of underlyings) produced a 48.9% win rate on subsequent options trades -- below the breakeven threshold and consistent with the academic finding that simple technical indicators do not reliably predict short-term options returns.

Highest-volume signals (buying the option with the highest daily volume in each underlying) produced a 46.1% win rate, worse than the baseline. High volume often reflects closing activity, hedging, and institutional roll trades -- not directional conviction.

Diamond signals outperform both alternatives by 5 to 8 percentage points of win rate, confirming that the scanner's multi-factor approach captures information that simple heuristics miss.

LIMITATIONS AND HONEST ASSESSMENT

Diamond signals are not infallible. Their 54.1% overall win rate means that nearly half of all diamond signals do not produce a profit within the 5-day window. The edge is real but modest at the individual signal level. It becomes meaningful through consistency across a large number of trades -- the law of large numbers working in the trader's favor.

Performance degrades in certain conditions: DRAINING regimes (for bullish signals), low-liquidity underlyings, and the first 15 minutes of the trading session (when flow data is noisiest). Traders who avoid these conditions can expect performance closer to the 57-58% range.

The backtest period includes a variety of market environments but may not capture all possible regimes. A sustained period of extremely low volatility, which did not occur in our sample, might produce different results.

CONCLUSIONS

Diamond signals outperform the broad options flow universe by 6.8 percentage points of win rate. The edge is statistically significant, economically meaningful, consistent across years, and robust to reasonable execution assumptions. Performance improves monotonically with confidence tier, is enhanced by regime filtering, and varies by underlying in predictable ways.

The most actionable takeaways: favor high-confidence and exceptional diamonds, align signal direction with the prevailing regime, focus on high-liquidity names where the edge is strongest, and use patient limit-order execution. These filters can push the effective win rate from the baseline 54.1% into the 60-65% range, which represents a substantial edge in options markets.

This analysis is retrospective. Past performance does not guarantee future results. Options trading involves substantial risk of loss. Diamond signal classifications are generated by iPresage's proprietary algorithms and may change over time.