DUK · Duke Energy Corporation
Stock Universe · mercado bursátil globalQuantitative Performance Analysis: Low-Frequency Multi-Week Algorithmic Strategy for DUK · DUK
This analytical report evaluates a top-ranked algorithmic trading strategy designed for DUK equity on a 15-minute timeframe over a 6.94-year historical testing window. Achieving a DevioLab score of 83.28 and holding the top rank for DUK, the strategy exhibits an exceptional 83.64% win rate across 55 closed trades, generating a cumulative return of 695.09% with a profit factor of 3.75. Characterized by high entry selectivity, positions are held for an average of 666.60 hours (roughly 27.8 days), bridging intraday bar evaluations with swing or positional holding durations. Risk parameters remain well-controlled, with a maximum historical drawdown of 13.24% that matches its worst individual trade loss (-13.24%) and a maximum consecutive losing streak of only 1 trade. Profit distribution is notably well-balanced, with the top three winning trades contributing just 25.38% of total gross profit. Recent performance since June 1, 2024, remains consistent, contributing 20 trades and 69.52% cumulative profit, confirming the ongoing structural stability of the strategy logic for execution via automated trading bots and trading agents.
Análisis al 21.10.2026. Los datos de la estrategia se actualizan por separado.Leer análisis completo
Strategy profile
The evaluated algorithm, designated internally as DUK 0 +695.09% 1TRAD-BON1, operates on 15-minute interval data for DUK stock across a comprehensive historical dataset spanning September 16, 2019, through August 24, 2026. Over this 6.94-year backtest window, the model completed 55 trades, ranking first among evaluated systems for DUK with an overall DevioLab score of 83.28. Designed as institutional-grade decision logic suitable for automated trading bots, trading agents, and execution algorithms, the system pairs high structural selectivity with extended trade durations. Across the entire testing duration, the system recorded 46 winning trades against only 9 losing trades, translating to an overall win rate of 83.64%. Cumulative closed-trade performance reached +695.09%, which corresponds to a normalized annualized return metric of 98.63%. The strategy maintains a profit factor of 3.75, signaling that gross profits generated by winning positions exceeded gross losses from unprofitable trades by nearly fourfold. Rather than seeking high turnover, the quantitative architecture emphasizes trade filtering, entering positions only when specific multi-timeframe conditions align on the 15-minute price series.
Trading rhythm and position duration
Despite sampling price behavior on a granular 15-minute timeframe, the strategy exhibits an ultra-low operational frequency that clearly distinguishes it from intraday scalping or rapid swing trading. The system averages 7.93 trades per year, or approximately 1.17 trades per active month. The statistical distribution of holding times reveals a multi-week position management model: average holding duration stands at 666.60 hours (approximately 27.78 calendar days), while the median holding duration is 480.25 hours (roughly 20.01 days). The difference between average and median holding times indicates that while most positions exit around three weeks into a trend, selective positions are held significantly longer to capture extended directional moves. Spacing between exits is similarly extended, averaging 44.15 days between closed trades, with a median exit interval of 33.97 days. This temporal rhythm demonstrates that an automated trading robot running this strategy would spend long periods either holding a single position through noise or sitting entirely in cash awaiting setup confirmation. For quantitative traders and automated execution setups, this pacing minimizes operational friction associated with hyperactive signal generation while capturing macro price developments in DUK.
Quality of historical results
Analyzing payoff metrics reveals that the high historical profit factor of 3.75 is supported by both solid win frequency and positive statistical expectancy on individual trades. The average return per completed trade is +4.14%, while the median trade yield stands at +3.41%. The closeness of the mean and median figures demonstrates that performance is not driven by extreme statistical noise or a single unrepresentative windfall. The strategy's best historical trade yielded +32.05%, whereas its worst individual trade resulted in a drawdown of -13.24%. Furthermore, examine the concentration of returns: the top three winning trades collectively generated 25.38% of total gross profit. This modest concentration percentage confirms that gross gains are broadly distributed across the 46 winning entries rather than relying on a tiny fraction of lucky outcomes. The structural asymmetry between average gains (+4.14%) and an 83.64% hit rate creates a robust mathematical profile where regular win accumulation easily absorbs occasional adverse exits.
Risk, drawdown and losing behavior
Risk metrics across the 6.94-year evaluation history exhibit exceptional containment. The maximum drawdown recorded across the entire backtest was 13.24%. Notably, this maximum drawdown figure matches the exact percentage magnitude of the worst individual losing trade (-13.24%). This statistical equivalence reveals a crucial aspect of the strategy's equity curve: peak-to-trough equity declines were primarily caused by single isolated adverse trade exits rather than compound, multi-trade loss sequences. This interpretation is reinforced by the losing streak statistics, where the maximum consecutive losing trades recorded was exactly 1. Across 55 historical trades, the algorithm never experienced back-to-back losses. With only 9 total losses occurring individually throughout the nearly seven-year backtest, equity recovery occurred immediately after every single loss. For automated trading agents programmed to execute this model, this behavioral profile reduces the psychological and operational strain typically caused by prolonged loss clusters or cascading drawdowns.
Behavior through time and yearly stability
A year-by-year examination reveals structural consistency across varying market regimes in DUK. Total trade count per calendar year remained tight, ranging from a minimum of 7 trades to a maximum of 10 trades per year. In 2020, the strategy logged 7 trades with 6 wins (85.71% win rate) and sum returns of +58.08%. In 2021, 7 trades produced 5 wins (71.43% win rate) totaling +35.60%. During 2022, 8 trades generated 7 wins (87.50% win rate) yielding +26.22%. In 2023, 10 trades delivered 8 wins (80.00% win rate) totaling +23.74%. In 2024, 9 trades yielded 8 wins (88.89% win rate) and +34.89%. In 2025, 7 trades yielded 6 wins (85.71% win rate) and +27.91%. Finally, in 2026, 7 trades recorded 6 wins (85.71% win rate) and +21.19%. Every single calendar year generated positive net cumulative gains while keeping win rates comfortably above 70%. The absence of negative years or sudden spikes in trade frequency indicates that the core entry and exit logic remained stable across multiple annual volatility cycles.
Recent period since 2024-06-01 versus full history
Evaluating recent behavior provides crucial insights into whether the algorithmic logic maintains efficacy in modern price environments. In the window from June 1, 2024, to August 24, 2026, the strategy closed 20 trades, generating a cumulative return of +69.52%. These 20 recent trades represent 36.36% of all trades completed over the 6.94-year history, reflecting a slightly higher signal frequency in recent years while maintaining the historical standard of execution quality. With an average return of approximately +3.48% per trade during this recent period, the recent trade returns align tightly with the full-history median trade return (+3.41%). This statistical alignment indicates that the performance properties observed over the broader backtest have not degraded in recent trading windows, offering empirical support for the system's continued operational stability.
Strengths and limitations
The primary analytical strength of this DUK strategy lies in its combination of a high win rate (83.64%), low drawdown (13.24%), and consistent historical profitability across every tested calendar year. The top three winners generating only 25.38% of gross profit highlights a dependable, non-outlier-dependent reward structure. Furthermore, an average holding time of ~27.8 days combined with low annual trade counts (7.93 trades/year) reduces exposure to short-term market noise. Conversely, the strategy's main limitation is its modest trade sample size. With only 55 completed trades across 6.94 years, individual statistical metrics carry broader confidence intervals than high-frequency systems with thousands of trades. Additionally, because average position holding times span several weeks, market participants utilizing automated trading bots must accept extended holding periods and overnight equity exposure, requiring sufficient capital buffer during multi-week position management.
DevioLab analytical conclusion
The quantitative evaluation confirms why this strategy achieves the #1 rank for DUK on DevioLab with a score of 83.28. By combining high signal selectivity on a 15-minute timeframe with multi-week position holding periods, the algorithm captures directional trends in DUK while avoiding overtrading. Its historical profile—characterized by a 3.75 profit factor, 83.64% win rate, single-loss maximum streak, and a modest 13.24% drawdown—demonstrates exceptional historical balance between profit capture and risk mitigation. The consistent multi-year performance and recent stability post-June 2024 reinforce its suitability as robust logic for automated trading robots and trading agents operating in equity markets.
Data scope and methodology
All metrics and statistics cited in this article are derived exclusively from historical backtested closed-trade performance for DUK on a 15-minute candle dataset between September 16, 2019, and August 24, 2026. The recent period statistics represent trades with closed timestamps on or after June 1, 2024. All return percentages reflect unleveraged, historical closed trade performance and do not include simulated slippage, brokerage commissions, or borrow fees unless explicitly stated. Historical simulated backtest results are provided for quantitative research purposes and do not guarantee future live execution performance.