Why Short-Horizon Forecasts Do Not Solve Trade Execution

Summary

The document considers whether a model that forecasts market behavior a few seconds ahead makes trading logic straightforward. The questioner describes fitting distributions to level-two order-book data and developing signals for both passive and aggressive strategies, then asks whether a stronger forecast permits simpler order-submission rules.

The replies emphasize that predictive quality is only one part of a trading system. A forecast must be defined in measurable terms before its practical value can be assessed, and converting it into orders still depends on trading logic, software reliability, error handling, and sound economic reasoning. The source offers no empirical results or specific execution algorithm, so it does not quantify how much forecast accuracy offsets implementation complexity. Its main lesson is that a short-horizon signal alone does not demonstrate monetizable performance.

Key ideas

• A short-horizon prediction needs a precise definition before its quality can be evaluated.
• Forecast accuracy is only one component of a functioning trading system.
• Order submission depends on execution rules, reliable software, error handling, and economic reasoning.
• The discussion gives no measured evidence that a strong signal makes execution logic trivial.

Tags

# Financial forecasting and Optimal order submission


# Financial forecasting and Optimal order submission












For instance, If i have a model that can accurately forecast 3s ahead, would the trading logic be rather trivial? I have fit a series of distributions to L2 data and believe I have a fairly good grasp as to what will happen in a few seconds. Now I am at the stage where I have to design some logic in order to get trades into the market. I am working on both passive and aggressive strat, as I have generated signals for both.

Just wanted to know if that if the better the model, then you can get away with having a more basic trading logic.

## Answer by wildbunny (score 1)

https://quant.stackexchange.com/a/43378

In my experience and good signal is only part of the problem you are solving. Good software engineering, good trading logic, good error handling and good economic reasoning all play important roles as well.

## Answer by Ezy (score 0)

https://quant.stackexchange.com/a/43385

Sorry for the trivial answer but first of all saying that you have a “good 3s signal” is not even a well defined statement, so i let you reassess your prior on the “triviality” of the monetization :)

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.