Energy Stock Rankings

Each stock receives a Composite Score (0–100) combining two independent analytical engines. Higher score = stronger combined signal. The score is relative — it ranks stocks against each other, not against an absolute threshold.

Composite Score

Weighted average of the Damodaran fundamental score and the Lo Adaptive signals score, normalized to 0–100. Currently 50/50 weighted. A score of 75+ indicates strong alignment across both engines. Scores are cross-sectional — a stock ranked #1 at 75 is the best in the universe today, not an absolute buy signal.

Regime  risk on transitional risk off

The current market state for this stock as detected by a Gaussian Hidden Markov Model (HMM) fitted on 2 years of daily log-returns and volume. The HMM identifies 3 latent states and labels them post-hoc by their return distribution:

• Risk On — positive mean return state; market is rewarding buyers. Momentum strategies tend to work here.
• Transitional — near-zero mean return state; regime is shifting or ambiguous. Reduce position sizing, wait for confirmation.
• Risk Off — negative mean return state; sellers in control. Even strong momentum names should be sized down.

Important: Regime is per-stock, not market-wide. A stock can be risk on during a broad market selloff if it is rotating into favor (e.g. defensive energy names during equity drawdowns).

Confidence

The posterior probability that the HMM assigns to the current regime state — how certain the model is about its regime classification. A regime label with <60% confidence is unreliable; treat it as transitional regardless of the label. High score + high confidence is the strongest combined signal. High score + low confidence (e.g. REMX at 42%) warrants caution even if the label says risk on.

Momentum

A composite of three signals, weighted 50/30/20:

• Cross-sectional momentum (50%) — rank-based z-score of 12-1 month returns across the universe. Captures which stocks are outperforming peers over the past year, skipping the most recent month to avoid short-term reversal. Based on Jegadeesh & Titman (1993).
• Trend alignment (30%) — price vs. 50-day and 200-day moving averages. Scores 0, 0.33, 0.67, or 1.0 based on how many of: price>MA50, price>MA200, MA50>MA200 (golden cross) are true.
• RSI normalized (20%) — 14-day Wilder RSI rescaled to [0,1] between oversold (30) and overbought (70). Not used as a contrarian signal here — higher RSI = stronger trend confirmation.

Kelly Size — Is This Relevant?

Yes — but treat it as a relative conviction signal, not a literal portfolio weight.

The Kelly Criterion (J.L. Kelly, 1956) is the mathematically optimal fraction of capital to bet on a wager with known odds and win probability, maximizing long-run geometric growth. Formula: f* = p/|loss| − q/gain

Here it is applied to stocks using:

• Win probability — fraction of positive daily returns over the past 63 days
• Expected gain — annualized mean return over 63 days
• Max loss — 2× annualized volatility (2-sigma drawdown proxy)

Three practical adjustments are applied:

• Quarter-Kelly (÷4) — full Kelly is theoretically optimal but brutal in practice; most practitioners use ¼ Kelly to reduce variance
• Regime multiplier — risk on: 1.0×, transitional: 0.5×, risk off: 0.25×
• 10% cap — no single name exceeds 10% of portfolio regardless of Kelly output

A Kelly of 0% means the raw Kelly formula returned zero or negative — the stock's return/risk profile doesn't justify a position at current prices even before regime adjustment. This is common for high-volatility names with inconsistent returns (UUUU, UEC, FLNC). High momentum + zero Kelly = interesting name but wait for a better entry.