AI Seasonality — SPY

Recalibrated seasonal path (TruePath) + AI Composite Score 1–10

What does this page show? — AI Seasonality combines classical seasonal analysis with machine learning. The TruePath finds the historically most similar years (correlation/DTW) and projects their weighted path into the future. The AI Composite Score (0–10) combines four sub-scores into a single signal: match quality, forward return, monthly win rate and tracking quality.
i
How to read the score: The KI Seasonal Score aggregates 4 independent sub-scores à 0–2.5 points into a total of 0–10.
Bullish ≥ 6.5 · Neutral 3.5–6.5 · Bearish ≤ 3.5

The higher the score, the more strongly historical seasonality, pattern match, and current trajectory point in the same direction. The score is not a trading recommendation, but a probability indicator for the seasonal setup over the coming weeks.
AI Seasonal Score
– / 10
–
Composite from 4 sub-scores
i
How to read the radar: Each axis shows one of the 4 sub-scores (0–2.5 points). The larger the golden area, the stronger the overall signal.

If the radar is one-sided — e.g. strong on Pattern path and Trend, but weak on Tracking — the current year is not yet following the expected pattern. A balanced, full radar means all four indicators point in the same direction.
TruePath vs. Current Path
Classical ∅ TruePath Current Year Match Years Projection

The pattern path recalibrates seasonality by combining only the historically most similar years in a weighted manner — as opposed to the classic average over all years. This filters out "noise" from structurally unsuitable years.

Most Similar Years (Top-N)
YearSimilarityAnnual ReturnPresidential Cycle
Methodology

TruePath (Pattern Path)

  • Normalize each annual curve to 100 (start) and interpolate to 365 calendar days
  • Calculate similarity between the current year (up to today) and each historical year
  • Pearson correlation: evaluates the shape of the curve (up/down patterns)
  • Euclidean distance: evaluates the absolute deviation of values
  • Select the top-N most similar years as "matches"
  • Weight each match year by its similarity and compute the weighted average
  • Smooth the result with a rolling window

AI Seasonal Score (1–10)

Composite of 4 sub-scores à 0–2.5 points:

  • Pattern path quality (0–2.5): How many of the similar years were positive?
  • Trend projection (0–2.5): Is the 30-day projection of the pattern path bullish or bearish?
  • Win rate (0–2.5): Historical win rate of the current month across all years
  • Tracking quality (0–2.5): How closely does the current year follow the seasonal pattern?

Signal: Bullish ≥ 6.5 · Neutral 3.5–6.5 · Bearish ≤ 3.5