Theta Forecast

The useful outcome

Theta is often described as a mysterious decomposition. The equivalent formula exposes a simple mechanism: a simple-exponential-smoothing level plus a carefully scaled drift. The practical goal is a path that another person can reproduce from a declared cutoff—not a line whose origin, state, or future assumptions are hidden.

The question this tutorial answers is: How does the classical non-seasonal Theta forecast combine a smoothed level with half the fitted linear trend?

What you will build

You will calculate the model by hand, implement it in Python and TypeScript, inspect eight long synthetic scenarios, and learn where correct recursion ends and empirical forecast validation begins.

Simple exponential smoothing and ordinary least-squares slope. Seasonal preprocessing is intentionally excluded.

Freeze the variant before fitting anything

Classical theta-zero/theta-two point forecast in its exact non-seasonal SES-with-drift form, fixed alpha in (0,1], OLS slope over t=0...n-1, and finite-sample correction.

This matters because forecast libraries often share a model-family name while differing in constants, trend representations, signs, initialization, missing data, or optimizer behavior.

Intuition before notation

The smoothed level carries recent information; half the global OLS slope contributes a deterministic drift with a finite-sample adjustment. The hardest part is the exact finite-sample drift correction and the distinction from an optimized drift model. The forecast origin is therefore a visible data boundary, not just the point where a chart changes color.

Formula and symbols

$\hat y_{T+h|T}=\tilde y_T+\frac{\hat b_0}{2}\left[h-1+\frac{1}{\alpha}-\frac{(1-\alpha)^n}{\alpha}\right]$

The formula is a conditional point forecast. Future random innovations use zero because zero is their conditional expectation—not because the future is known to contain no shocks.

Data that the formula is allowed to see

A finite non-seasonal equally spaced series, fixed SES alpha, and positive horizon. No missing values or automatic seasonality test. Observations must be equally spaced, finite, ordered, and on one consistent scale. The model version and cutoff are frozen before the holdout. Revisions require a new forecast version rather than silent history replacement.

Algorithm

1. Validate the non-seasonal series, alpha in (0,1], and horizon.
2. Fit the ordinary least-squares time slope using indexes 0 through n-1.
3. Run simple exponential smoothing with the first observation as initial level.
4. Calculate the finite-sample correction from alpha and n.
5. Add half the OLS slope times the horizon correction to the final SES level.

Hand calculation

The following is author-derived arithmetic over synthetic values:

{
  "values": [
    10,
    12,
    14,
    16
  ],
  "config": {
    "alpha": 0.5,
    "horizon": 2
  }
}

• OLS slope b0=2; final SES level at alpha=.5 is 14.25.
• Finite-sample base is 1/.5-(.5^4)/.5=1.875.
• h=1: 14.25+1(1.875)=16.125; h=2 adds 2.875, giving 17.125.

The expected path is [16.125, 17.125]. Independent tests pin these values before the larger fixtures are generated.

Calculation flow

The order matters. Revealed actuals may score the path, but they may not alter the version that was issued at the cutoff.

Guided experiment

The standalone lab contains eight scenarios and 1,440 calendar-month-end observations. It trains on 144 values and forecasts a 36-value holdout. Start at horizon one, predict what state should persist, then Step through twelve model-specific checkpoints. Change the principal parameter and watch the complete path recompute from the frozen training data.

The scenarios isolate persistence, mean reversion, linear trend, seasonal structure, a late level shift, one gross outlier, a variance-regime change, and seasonal phase/amplitude drift. They are not selected trades or provider histories.

The state an expert audits

For this model, keep the final SES level, OLS slope, finite-sample correction, and horizon-specific drift contribution visible.

Use classical nonseasonal Theta as a strong transparent baseline when a smoothed level plus constrained linear drift matches the teaching or operational question. Handle seasonality explicitly before applying this selected variant.

Separate the SES level from the drift contribution. Endpoint or structural-break sensitivity appears mainly through the global OLS slope; long horizons extend that slope mechanically.

When implementations disagree, use this order: Check deseasonalization, alpha estimation versus a supplied alpha, time-index origin, finite-sample correction, OLS slope definition, and whether the implementation uses the classical or generalized Theta method.

The workbench is deliberately quantitative. It connects the model's unique state to the declared benchmark rather than asking the learner to judge whether the purple line “looks right.”

Failure diagnosis

• Automatic seasonality detection and deseasonalization are outside this selected variant.
• Alpha equal to zero is invalid because the exact correction divides by alpha.
• A global OLS slope is sensitive to level shifts and endpoints.
• Large h extends deterministic drift indefinitely.
• The method's historical competition results do not guarantee performance on a new financial series.

Invoking the M3 result as evidence of universal superiority or omitting the seasonal preprocessing difference. Correct arithmetic answers “did we implement the selected model?” It does not answer “is this a good model for this series?”

Implementation contract

The Python reference favors readable state transitions. TypeScript implements the same contract independently. Shared fixtures and focused tests require matching numeric paths within 1e-9 and exact agreement on lengths, nulls, states, and invalid-input behavior.

Estimation is intentionally outside the public function. In a real workflow, fit or select parameters using only training data, record the method and version, pass the frozen result into this recursion engine, and evaluate using rolling origins against simple benchmarks.

Practical evaluation

Use a loss function that matches the decision, report by horizon, and include baseline forecasts. Preserve the refit schedule and origin timestamps. If prediction intervals are needed, choose and validate an uncertainty model; do not manufacture intervals from point forecasts alone.

The lab's declared baseline is nonseasonal naïve last-value forecast. At active horizon k, it calculates errors only for the first k revealed holdout values:

• MAE = mean(abs(actual - forecast));
• RMSE = sqrt(mean((actual - forecast)^2));
• bias = mean(actual - forecast);
• MAE skill = 1 - model_MAE / benchmark_MAE.

Positive skill means the model's revealed-prefix MAE is lower than the frozen benchmark's. Negative skill is useful evidence: the sophisticated recursion is losing to a simple rule on this synthetic path. When benchmark MAE is zero, skill is undefined rather than infinite. One origin still cannot establish general performance; repeat the complete procedure over rolling origins before making an empirical claim.

Related methods

AutoReg introduces lag memory. ARMA adds innovation memory. ARIMA changes the modeled scale through differencing. SARIMA/SARIMAX adds seasonal polynomials and future driver scenarios. Holt-Winters maintains level, trend, and seasonal states directly. Theta combines a smoothed level with constrained drift.

Next: Vector Autoregression.

Historical and evidence boundary

A historical example is not useful for this mechanism lesson. A defensible case requires licensed point-in-time observations, identity, frequency, adjustment and revision basis, cutoff, parameters frozen before inspection, benchmark, holdout, and publication rights.

This tutorial establishes definition fidelity. It does not establish stationarity, model adequacy, forecast calibration, historical association, causation, profitability, or investment suitability.

Primary references

• The theta model: a decomposition approach to forecasting — Theta lines modify local curvature and the classical method combines theta-zero and theta-two forecasts.
• Unmasking the Theta method — The classical point forecast is equivalent to simple exponential smoothing with a drift equal to half the fitted OLS slope, including a finite-sample correction.
• The Theta Model — Theta forecasting combines OLS drift and simple exponential smoothing, with optional seasonality handling.
• Forecasting: Principles and Practice — Evaluating point forecast accuracy — Forecast accuracy must be measured on genuine forecasts outside the training sample; MAE, RMSE, scaled errors, and benchmarks answer different evaluation questions.
• Forecasting: Principles and Practice — Time series cross-validation — Rolling forecasting origins preserve temporal order and can evaluate horizon-specific errors without using future observations.
• Forecasting: Principles and Practice — Evaluating regression and residual behavior — Residual plots and residual autocorrelation diagnose information left in the fitted model, but residual size is not a substitute for genuine forecast error.