Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
94 tokens/sec
Gemini 2.5 Pro Premium
55 tokens/sec
GPT-5 Medium
38 tokens/sec
GPT-5 High Premium
24 tokens/sec
GPT-4o
106 tokens/sec
DeepSeek R1 via Azure Premium
98 tokens/sec
GPT OSS 120B via Groq Premium
518 tokens/sec
Kimi K2 via Groq Premium
188 tokens/sec
2000 character limit reached

Breaking the Trend: How to Avoid Cherry-Picked Signals (2504.10914v2)

Published 15 Apr 2025 in q-fin.PM

Abstract: Our empirical results, illustrated in Fig.5, show an impressive fit with the pretty complex theoritical Sharpe formula of a Trend following strategy depending on the parameter of the signal, which was derived by Grebenkov and Serror (2014). That empirical fit convinces us that a mean-reversion process with only one time scale is enough to model, in a pretty precise way, the reality of the trend-following mechanism at the average scale of CTAs and as a consequence, using only one simple EMA, appears optimal to capture the trend. As a consequence, using a complex basket of different complex indicators as signal, do not seem to be so rational or optimal and exposes to the risk of cherry-picking.

Summary

Breaking the Trend: How to Avoid Cherry-Picked Signals

The paper "Breaking the Trend: How to Avoid Cherry-Picked Signals" by Sebastien Valeyre provides a critical examination of trend-following strategies within the context of Commodity Trading Advisors (CTAs). The author challenges the conventional wisdom of using a blend of various technical indicators, suggesting that a simpler approach using a single Exponential Moving Average (EMA) may be more effective. This paper leverages the theoretical insights provided by Grebenkov's model to empirically validate the utility of EMA in trend-following strategies.

Empirical Validation of Grebenkov's Model

Valeyre focuses on the optimization problem tackled by Grebenkov, highlighting an elegant formula for the theoretical Sharpe ratio of trend-following strategies. Grebenkov's work suggests that portfolio positions should linearly depend on signals, specifically on the EMA of returns. This diverges from the traditional binary choice of long/short positions. The empirical results presented in the paper align closely with Grebenkov's theoretical predictions, using parameters such as λ=1180\lambda = \frac{1}{180} and β0=0.12\beta_0 = 0.12 to achieve an optimal Sharpe ratio. The author finds that a single EMA using a 112-day time scale is optimal, challenging traditional methods that employ multiple complex indicators.

Implications for Trend-Following Strategy

Valeyre's analysis suggests that a simplified strategy using a single time-scale EMA is not only adequate but preferable. This is a salient point given the inherent complexity and potential pitfalls of using multiple indicators, such as exposure to cherry-picking. The findings implicate that market trends can be sufficiently captured by modeling them as mean-reversion processes with a singular time scale, rather than relying on multi-time scale approaches.

Critique of Multi-Indicator Approaches

The paper critiques the juxtaposition of complex indicator suites against the simplicity of the EMA. Valeyre posits that using a complex arrangement of indicators poses unnecessary risks, given the robust performance of a singular EMA. By validating Grebenkov's theoretical model with empirical data, the paper highlights that existing market strategies might overcomplicate trend-following to their detriment. The emphasis on simplicity could lead CTAs to reevaluate their signal construction methodologies.

Potential for Future Research

While the paper strongly supports the use of a single EMA, the potential role of shorter time scales is acknowledged, although deemed significantly less impactful on medium-frequency strategies. This raises questions about the practical application of short-term signals and calls for further investigation into their specific utility within a broader trading strategy framework. Additionally, the strong correlation between indicators with varying parameters invites exploration into underlying factors that drive these correlations, which might point to common market influences or psychological factors in trading behavior.

Conclusion

Valeyre's paper makes a compelling case for rethinking trend-following strategies, encouraging a pivot towards simplicity and rigorous empirical validation over traditional diversity of signals. By corroborating Grebenkov's model with empirical evidence, Valeyre demonstrates the efficacy and elegance of employing a singular EMA-based approach, offering a streamlined alternative to signal construction for CTAs. As a result, the paper sets the stage for further exploration into optimizing trend-following methodologies within financial markets.

Dice Question Streamline Icon: https://streamlinehq.com

Follow-up Questions

We haven't generated follow-up questions for this paper yet.

Authors (1)

Don't miss out on important new AI/ML research

See which papers are being discussed right now on X, Reddit, and more:

“Emergent Mind helps me see which AI papers have caught fire online.”

Philip

Philip

Creator, AI Explained on YouTube