Learn / Topic cluster
Topic cluster
16 answers
Signal calibration and thresholds
A cluster on probability calibration, confidence-to-size mapping, hysteresis, and the filters that separate a promising score from a live trading rule.
A model score is not useful just because it looks smooth or predictive. Someone still has to decide what the score means, whether it is calibrated enough to trust, where the trade threshold belongs, and when a seemingly strong forecast is too weak to survive real costs.
This cluster connects those decisions back to Alphora's broader research loop. It treats calibration, thresholding, and no-trade logic as explicit design choices that should survive walk-forward validation, portfolio context, and live-paper discipline instead of hiding inside a single magic cutoff.
Questions in this cluster
Each page answers a narrower search-shaped question while staying linked to the broader research theme.
Strategy intuition
definition
What are calibration curves?
calibration curves is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
Why do calibration curves matter in systematic trading?
calibration curves is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
What are hysteresis bands?
hysteresis bands is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
Why do hysteresis bands matter in systematic trading?
hysteresis bands is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
What is confidence decay?
confidence decay is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
Why does confidence decay matter in systematic trading?
confidence decay is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
What are decision thresholds?
decision thresholds is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Strategy intuition
definition
Why do decision thresholds matter in systematic trading?
decision thresholds is one of the core ideas inside signal calibration and thresholds. It matters because it changes how a researcher turns a clean intuition into a repeatable rule about selection, sizing, timing, or validation.
Research process
implementation
How do you calibrate a trading probability?
Calibrating a trading probability means making the forecast line up with realized outcomes or net expectancy well enough that a stated confidence level actually means something after costs and delay.
Research process
research
Why can a strong signal still be untradeable?
A signal can look statistically strong and still be untradeable when the edge is too small, too slow, too crowded, or too expensive to express under real execution, sizing, and capacity constraints.
Research process
implementation
How do hysteresis bands differ from simple thresholds?
A simple threshold flips the strategy at one line. A hysteresis band uses different entry and exit conditions or a state-dependent deadband, so the signal has to move meaningfully before the position changes again.
Risk management
implementation
How should you map model confidence to position size?
Confidence should influence size only after the score is calibrated and filtered through risk, liquidity, correlation, and cost constraints. Raw model conviction is not the same thing as safe leverage.
Research process
implementation
How should you set a trading threshold when costs are nonlinear?
When costs rise with size, volatility, or thin liquidity, the threshold should clear a moving net-value hurdle rather than a fixed raw score cutoff.
Research process
research
Why does a trading probability look calibrated in sample but fail live?
A trading probability can look well calibrated in sample and then fail live because the regime, label process, data timing, execution conditions, or class balance changed in ways the calibration layer did not learn to handle.
Feature intuition
definition
When should you binarize a continuous trading score?
Binarize a continuous score when the useful edge mostly lives in the tails or when execution only supports sparse, high-conviction trades. Keep it continuous when intermediate values still carry ranking or sizing information that the portfolio can use.
Research process
implementation
How do you walk-forward test a calibration or threshold rule?
Walk-forward testing a calibration or threshold rule means refitting it only on the training slice, freezing it for the next forward window, and checking whether reliability, trade count, and net performance remain stable as time moves on.