In multi-horizon forecasting, the short horizon lies first
Tetracta AI Teams · 15 June 2026
A methodology note from our quantitative-research practice. It describes a general lesson about overfitting and honest backtesting. No strategy, instrument, universe, dataset, or performance figure is disclosed.
When you forecast the same target at several horizons — short, medium, long — there is a seductive instinct to chase the short one. It feels like the easy money: more observations per unit of calendar time, faster feedback, quicker iteration. In our experience it is usually where you fool yourself first.
Why the short horizon overfits
At short horizons the signal-to-noise ratio is brutal. Most of what moves a short-horizon outcome is noise — timing, microstructure, things no model should be able to predict from the features it has. A flexible model handed a short horizon and limited data will happily memorize that noise and report a beautiful in-sample fit. The shortness is the trap: there is so little true signal that almost all of the apparent skill is the model reciting what it has already seen.
The longer horizon is less glamorous and more honest. Slower-moving structure is harder to overfit and likelier to reflect something real. In our own multi-horizon work the pattern was stark: only a longer horizon carried a signal that survived honest testing; the shortest horizons were, to a first approximation, noise that a model could memorize but not generalize.
The discipline that catches it
The danger is that a short-horizon overfit looks identical to a real edge until it meets reality. Four practices, in order of importance, separate the two.
- Leak-free, point-in-time backtesting. The most common way short-horizon results get inflated is lookahead — using a feature that, in production, would not yet be known at decision time. Short horizons are exquisitely sensitive to this; a sliver of accidental future information is enough to fabricate an edge. Every feature must be stamped with the moment it was actually available, and tested as of that moment.
- The temporal placebo. Shift or shuffle your labels in time and re-run the entire pipeline. If the "signal" survives a placebo where it cannot exist, you have not found an edge — you have found a leak or a bug. This single test has caught more false positives for us than any other.
- Out-of-sample across regimes. A result that only holds in the window you developed it on is a memory, not a model. Multi-year, walk-forward, across different conditions — or it does not count.
- Beware recency-overfit. Retraining constantly on the freshest data feels responsive; often it just chases the latest noise. We have watched "improvements" that were nothing but a model overfitting the recent past and degrading the moment the present stopped resembling it.
And measure honestly
A high t-statistic is not the same as a realised risk-adjusted return; an impressive gross number is not a net one. The horizon that overfits is also the one where costs bite hardest, because you act on it more often — a short-horizon result that looks wonderful gross can be flat or negative once you pay to trade it. If the figure you quote is not the one you would actually live on, it is decoration.
Why this generalizes
This is not really a finance lesson; it is an overfitting lesson that finance happens to make vivid. Anywhere you predict at multiple timescales with limited data — demand, load, equipment failure, churn — the shortest horizon is where leakage and overfit most convincingly masquerade as skill, and where a small team without an army of validators is most exposed. The defense is the same everywhere: assume the short horizon is lying until a placebo, a leak-free backtest, and an out-of-sample run across regimes prove otherwise.
We keep the specifics of our own systems closed — the targets, the features, the universe, the numbers. The discipline above, though, is not proprietary, and it is the part worth sharing: the cheapest edge in the world is not overfitting in the first place.
— Tetracta AI Teams · for humans, like humans.