Trading Unproven Signals and Expecting Positive Average Returns
Summary
The document asks whether traders can systematically act on a broad class of signals, including unproven patterns, while managing risk and expecting false signals to offset one another. It frames this as a question about the trade-off between mistakenly treating noise as meaningful and overlooking genuine opportunities, and asks how to choose a confidence threshold and backtest such an approach.
The brief responses offer only a limited perspective. One argues that quantitative strategies already work this way in a broad sense: individual signal instances need not succeed if the strategy has positive average returns. Another gives an anecdote about a trader using astrology, without evidence or methodological detail. The discussion does not provide a formal framework for signal selection, a method for setting confidence levels, or a backtesting procedure. Its useful takeaway is the distinction between single-trade accuracy and aggregate expectancy, while the anecdote does not establish that arbitrary or superstitious signals are profitable.
Key ideas
- A signal-based strategy can tolerate losing instances if its average returns are positive.
- The question proposes trading a broad class of signals while controlling capital and risk.
- Choosing a confidence threshold for acting on a signal is raised but not answered.
- The responses provide no systematic backtesting method or evidence that false signals reliably cancel out.
- An anecdote about astrology is not evidence of a profitable trading method.
Tags
Full text
# Is there something like opportunistic "superstitious" trading? # Is there something like opportunistic "superstitious" trading? This sentence in the following paper got me thinking: "Some traders [...] trade every pattern whether proven or not, expecting authentic ones to produce positive results, whilst the profits and losses of fake patterns cancel each other out." http://www.seasonalcharts.com/img/ZUTEXTEN/saisonalitaet_e.pdf Then I got my hands on another paper - seemingly unrelated...: "The evolution of superstitious and superstition-like behaviour": http://rspb.royalsocietypublishing.org/content/276/1654/31.full.pdf The bottom line is that I can be rational to act irrational, or in other words there is a trade-off between being superstitious and being ignorant - and being superstitious (i.e. seeing patterns where there are none) can under certain circumstances be beneficial - this is why it survived evolution up until now. Have you come across research that systematized this approach for the trading arena, i.e. first: what is the right confidence level to trade a signal and second "opportunistic trading" as an approach of its own? With "opportunistic trading" I mean a framework to trade every signal out of some class (with money and risk management attached for not going bust) in the hope that the false signals cancel each other out and the real ones make money. Perhaps you have some thoughts how to backtest these ideas, too. ## Answer by AnonQuant (score 4, accepted) https://quant.stackexchange.com/a/384 In a sense all quantitative strategies do that because no one expects every instance of a signal to create a positive return. The expectation is that on average it will. Recall that these questions are supposed to be about quantitative finance whereas the first link you have seems to be related to technical analysis. ## Answer by glyphard (score 2) https://quant.stackexchange.com/a/289 Not exactly definitive, but I once met a lady at a seminar I was teaching that gave an overview of how she uses astrology to make trading decisions.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.