← sources

Researching NQ changed how I think about trading edge

source: reddit (reddit:algotrading)published 2026-10-31 01:28ingested 2026-11-01 02:41

retained text

English is my second language, so I used GPT to help with the translation.

I have been investing in stocks for more than 20 years. Day trading came much later, as a hobby I picked up toward the end of the year. I currently run one small Express account. After spending so many years investing in stocks, I wanted to develop a more systematic way to make money in falling markets as well. 

My basic approach has always been to build knowledge about the market. I collect data, study its behavior, and organize the findings so I can return to them and test new ideas. Over time, that collection of market data, analysis, and research material has grown enough to fill a 40TB HDD.

Through this work, I became interested in two connected problems: where a strategy’s profit actually comes from, and how to judge whether that source of profit is still there.

**1. Understanding where the profit comes from**

My early research focused on relatively simple trend strategies. In some cases, I spent a month developing and testing an approach and still struggled to reach PF 1.0 after costs. That is not a general criticism of trend following. It describes the approaches I built, under the conditions I tested.

The strategies that reached PF 2 or higher in my backtests generally captured more specific behavior within the market’s movement. Broad direction alone did not explain the results. What mattered was the path after entry: **how far price moved in our favor, how much it moved against us, and whether the favorable move arrived before the stop or time limit.**

For example, our research included profitable short trades during a rising session. The system did not need to predict a bearish day. It needed to identify a temporary downward move and exit while that opportunity was available.

Examining the exits was especially useful. In a diagnostic using the same entries, removing the profit targets caused several winners to become losses as price later reversed. Other trades improved without the target because the favorable move continued.

That showed both sides of the decision. A target can preserve a temporary profit, but it can also cut off a larger winner.

My conclusion was that the entry condition and exit policy had to be understood together. The profitable behavior was not simply “high volatility.” More volatility can bring the stop closer in time just as easily as the target.

In my better results, the useful relationship was between a particular market condition, the price movement that followed, and a trading policy suited to that movement.

**2. Evaluating whether that relationship is still valid**

This created the next problem.

**If a strategy depends on a particular pattern of price movement, its performance will change when that pattern changes.**

Moves that previously continued may start reversing. Favorable excursions may become smaller. Adverse excursions may become deeper. The entry signal can remain the same while the distribution of outcomes changes.

**That is the edge decay I wanted to address.**

The difficulty is distinguishing it from ordinary drawdown. Recent losses do not necessarily mean a strategy has stopped working. Recent profits do not necessarily mean it still has an edge. There is also the possibility that the original result was mostly overfitting.

A recent PF number cannot resolve those questions by itself.

So I built a system to evaluate whether there is still sufficient evidence to keep using a strategy. The focus is on whether the behavior behind its profitability continues to appear, whether performance depends too heavily on a few trades, whether its loss characteristics have changed, and whether it remains economically useful after costs.

I am not trying to identify the exact moment an edge disappears. I want a consistent basis for deciding whether a strategy still deserves to be traded.

Today marks the beginning of the second week with my current setup. The first week performed better than I expected. I do not expect the initial results to continue unchanged. My expectation is that maintaining a net PF above 1.5 may be achievable, but that remains something to verify over a longer period.

This research changed my objective. **Finding a profitable strategy is still important, but I no longer see a strong historical backtest as the end of the work.**

document fetched 2026-11-01 00:40 · canonical https://reddit.com/r/algotrading/comments/1xjxf88/researching_nq_changed_how_i_think_about_trading