Why Trading Can Erase Predictable Market Patterns
Summary
The answer argues that exploiting a predictable market pattern can change prices and weaken the pattern itself. It illustrates this with a hypothetical weekday effect: if traders buy ahead of the expected rise and sell into it, their activity may shift prices until the original signal no longer offers the same opportunity. This is an intuitive account of how competition and trading affect market efficiency.
The document describes possible exceptions where demand is constrained by regulation, accounting needs, or limited liquidity, and where a large trade can disrupt prices. It cautions that market behavior differs from a physical resonant system because participants adapt and mechanisms change. These points are conceptual rather than a tested model: no data or quantitative evidence is presented, and the broad claim that analysis enforces randomness should not be read as a universal law. Persistent effects may depend on costs, capacity, and market conditions.
Key ideas
- Trading on a predictable pattern can alter prices and erode the pattern's profitability.
- Regulatory or accounting flows may create temporary demand that is difficult for other traders to offset.
- Large trades and limited liquidity can make markets less efficient in particular situations.
- Financial participants adapt, so market patterns may be unstable compared with physical systems.
- The answer offers intuition rather than empirical tests or a general quantitative model.
Tags
Full text
# Is it possible that some types of financial systems can resonate? # Is it possible that some types of financial systems can resonate? Financial systems can certainly be modeled using the same tools physicists use to model dynamic physical systems. The validity of such is evidenced by models such as that developed by Black and Scholes to predict market outcomes. And I have first hand knowledge that engineers and physicists are hired by Wall Street to develop and apply such models to gain an edge on investment. So my question, from a physicist's or mathematician's point of view, who may have worked with financial systems - are there any financial models that can lead to any type of resonance - such as for example the flow of cash behaving like a standing wave? ## Answer by Phil H (score 3, accepted) https://quant.stackexchange.com/a/16346 The general effect of quantitative analysis of the markets is to enforce randomness. Suppose a strategic quant finds a predictable pattern where a stock always rises on Tuesdays. His institution will commence buying the stock every Monday, and selling on Tuesday. The trading itself pushes the stock price up on Monday and down on Tuesday (in general), so if enough volume is traded the pattern will disappear. The occasions where this is not true are when there are unavoidable demands due to regulation or when the market is not efficient; at the end of the year there is generally a strong demand for USD to satisfy regulatory or accounting requirements for capital, which push exchange rates and interest rates one way for a short period. It is largely predictable, but only those who are not short of USD can benefit; the pool is too small to correct the imbalance. Similarly, when a whale trades (e.g. the London Whale), the market is not efficient and other factors like liquidity and trade volume come into play. So the short answer is that any pattern like a resonance pattern should be eliminated in a fast, efficient market, and that there are enormous resources devoted to locating patterns in the numbers in order to exploit and effectively eliminate them. In a physical system, the particles are blind and dumb, and the mechanisms are predictable and consistent. In finance, every participant has complex behaviour, and many of the mechanisms are neither predictable nor consistent, except to consistently act to eliminate predictability. The history books are full of clever quants who had a sure thing until they didn't. For example, LTCM, the whole MBS setup, etc. The exception is Goldman.
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.