A Predictive Stock Entry & Exit SystemBuilt on OVTLYR behavioral analytics

Headline

The system identifies the current market regime (Crisis / Normal / Extended) from SPY breadth, scores each stock with a regime-aware composite, and exits each trade with five profit and deadline rules (T1/T2/T3/T3_EXT/CAP) plus two risk rules: a disaster stop and a dead-trade exit. Each position is sized so that hitting its stop costs 1% of the account, and half that on days with weak buying conditions. Idle cash sits in SPY while the Balancer strategy is long SPY, and in T-bills (SGOV) otherwise. Universe filters (average volume ≥ 1,000,000 and price ≥ $5) keep illiquid micro-caps and reverse-split-to-death stocks out of the candidate list.

Per-trade results — four validation slices

500 tickers were held out of all training and rule selection. Combined with a time split at 2024-01-01, this gives four test slices. Slice C (unseen tickers × unseen time) is the strictest test: tickers the rules were never tuned on, over a period the rules were never fitted to. Only completed trades are counted, and the $5 price screen uses the close on the entry day.

Win rate stays between 76% and 83% on every slice. The two recent slices are lower: 2024-2026 has been a harder market for fear entries than 2020-2023, which held two bear markets.

How much of this depends on which tickers were held out? Split the holdout's tickers into four disjoint quarters and you get four separate reads of the same rules over the same period:

Across 2,000 random half-samples the win rate has a 10th–90th percentile range of 74.0–78.2%, and the mean return runs from +1.03% to +6.76%. That's why the headline is a range, not a single number.

How trades end

With the live rules (the stop and the dead-trade exit included, every fill at the next open), on the holdout tickers, 2020-2026, 646 completed trades:

About three trades in four end at a profit exit. The profit exits fire only when the trade is in profit, so they win by construction; a few lose a little when the next open gaps below. The losing quarter mostly ends through the dead-trade exit, around day 60 at about −11%, and only 1 in 20 trades still reaches the 120-day cap. Each position is sized so a stop costs about 1% of the account, so a −29% stop exit costs the account about 1%.

The headline figures above are the entry and exit rules alone. Without the risk rules, 24.9% of trades on Slice C reach the cap, with an average loss of 23%.

Win rate by validation slice

Train (n=5,093)

82.74%

Slice A — unseen tickers (n=1,168)

82.62%

Slice B — train tickers × 2024+ (n=2,551)

77.58%

Slice C — both unseen (n=639)

76.06%

Random baseline (~53% drift)

53%

What is OVTLYR?

OVTLYR (pronounced "outlier") is a behavioral analytics platform that quantifies investor sentiment for 3,300+ US-listed stocks. Its core output is the Fear & Greed oscillator — a 0–100 score per stock per day. Lower means more fear, higher means more greed.

The system on this page uses sentiment as the primary feature, with price-based signals (EMA deviations, sector breadth, order blocks) as confirmations.

Metrics are documented in the glossary.

Methodology

The pipeline

1. Feature engineering. Extract Fear & Greed (F&G), EMA50 deviations, sector breadth and F&G rate of change across all 11 GICS sectors.
2. Bottom identification. Find ~7,000 price bottoms using local-minima detection plus forward-return analysis.
3. Stock classification. Label each ticker Mean-Reversion (MR), Momentum (MOM), or Mixed (MIX) from how it has responded to fear signals, using only data before each date.
4. Cross-ticker pattern mining. Find features that predict profitable entries regardless of ticker.
5. Model building. Build a scoring model for each regime.
6. Backtesting. 2020–2023 in-sample (IS), 2024–2026 out-of-sample (OOS). After each trade, the next entry on that ticker can only come after the exit, so overlapping holds are never counted twice.
7. Consolidation and CV. Cross-validate any candidate rule across year, regime and classification splits before shipping.
8. Portfolio testing. Run the rules as a 20-position portfolio with real fills at the next open and trading costs, and judge changes on return, drawdown and Sharpe in both time halves.

Scale

• 3,300+ stocks in the OVTLYR universe. Universe filters: volume ≥ 1M + price ≥ $5.
• 500 tickers held out of all training and rule selection.
• 11 GICS sectors represented.
• 4,358 out-of-sample (OOS) trades across Slices A, B and C, plus 5,093 in-sample training trades.
• Training: Jan 2020 – Dec 2023. Time holdout: Jan 2024 onward.
• Rule discovery: ~1,000 candidate rules tested with BH-FDR multiple-testing correction (q=0.10).

The Three Regimes

The market is split into three modes — Crisis, Normal, and Extended — based on how broad the market's uptrend is. Each mode has its own buy criteria. The mode is detected from SPY breadth, the percentage of S&P 500 stocks currently in uptrends.

SPY breadth — the share of S&P 500 stocks in an uptrend

Crisis

Normal

Extended

0%15%50%100%

Crisis0–15%

Deep fear. Low F&G, sentiment still falling, price well under the 50-day average, weak sector.

~30 days a year · highest conviction

Normal15–50%

Stock-level fear starting to turn. Below-neutral F&G ticking up, price under the 50-day average.

Most of the year · the everyday playbook

Extended50–100%

No new entries. Over half the index is already in an uptrend, so this is late in the move — the weakest forward returns of any regime. Existing positions are held.

16% of days · no entries taken

The strategy inverts across modes. In Crisis, panic-selling creates the best buying opportunities — everyone wants out, the stock is cheap, the recovery edge is largest. The mode is always checked first; everything downstream depends on it.

The n810 Trading Plan

The entry and exit side has three pieces: universe filters (volume ≥ 1M, price ≥ $5), the entry filter (F&G < 70 in every regime), and five exit rules. Two refinements: T1 is skipped while OVTLYR's trend overlay is up, and T3_EXT is skipped for MOM stocks so they can run past their first peak. See stock classification for the MR/MOM/MIX definition.

Universe filters (applied before any other rule)

The five exit rules

Each day after entry, these rules are checked in order. The first one that matches sells the position at the next open. Four of them require the trade to be profitable; the fifth (CAP) is a hard 120-day deadline.

Each day a position is open, in this order

1. 1T1Sentiment back in greed, trade up 5%+ — unless the trend is still confirmed→ sell
2. 2T2Entered in deep fear, the bounce has faded, trade up 5%+→ sell
3. 3T3Opened in Crisis, sentiment recovered to 40, trade up 5%+→ sell
4. 4T3_EXTOpened outside Crisis, peaked at +10%, now fading — momentum names exempt→ sell
5. 5CAPNone of the above fired within 120 trading days→ sell

↺No match — hold, and check the same five tomorrow.

T3_EXT usually sells at or near the trade's peak: the median give-back from the peak is zero. After the sale, on the train tickers, the stocks did no better than SPY over the next 60 days.

How to Trade It

The complete rule set, as the daily picks publish it. Every rule is in one shared module that the picks, the position checker and the backtests all call.

Every morning (picks publish before the US open):

1. Buy new picks at the open. Each pick carries a stop price and a size.
2. Size each position so that hitting its stop loses 1% of the account: 1% ÷ the stop distance, capped at 8% of the account. A typical pick is about 3%; a volatile one is smaller, a calm one larger. Hold at most 20 positions. To risk 0.5% or 2% per position instead, scale every size by the same factor. On Weak buying-condition days, new picks get half that size. The picks show the halved size. Weak-day trades still make money on average, just less, so they are bought smaller, not skipped.
3. Sell at the open when a position meets any exit:
   • the profit exits T1, T2, T3, T3_EXT, or the 120-day CAP;
   • Dead trade: held 45+ trading days, never closed 5%+ above the entry price, below the entry price today, and OVTLYR has printed 2+ Sell signals since you bought;
   • Disaster stop: a close at or below the entry price − 8 × ATR20, where ATR20 is the stock's average daily range over the 20 days before you bought. That's about −33% for a typical pick. It catches disasters, not dips. An overnight gap can still open below it, which is why the size is set from it.
4. Idle cash (anything not in positions) goes into SPY while the Balancer strategy is long SPY, and into SGOV (0–3 month T-bills) otherwise. The picks page shows which every morning. When it changes, act that day, at the open or the close. Sell SPY or SGOV as needed to fund new picks.

Performance

By year (Slice C: unseen tickers × 2024-2026)

Trades opened in 2026 are mostly still open, and a trade only counts once it has closed, so 2026 is not shown.

Why sector breadth carries more weight at its low tail

The entry score reads sector breadth: the share of the stock's own sector on an OVTLYR Buy signal. The lower it is, the better the trade, all the way down to near zero, on all four slices. So the score grades the tail: +1.5 below 2%, +1.0 below 5%, +0.5 below 10%, on top of the standard breadth points. Inside a single regime the effect holds, and it survives a control for the stock's own Fear & Greed.

Market and sector conditions

Weaker sector and market conditions at entry go with better trades, in both regimes. For example, trades opened while the stock's sector ETF was below its 50-day average returned +5.3% on average, against +1.1% when it was above. Trade results are still about 90% stock-specific: the market explains roughly a fifth of a single stock's moves, and the sector adds a few points more.

Portfolio Results

The tables above are per trade. This is the full rule set run as a portfolio: 2020-01 to 2026-10, every fill at the next open, 10 bp trading costs, 20 positions, $100k start. Rule-set rows are the median of 20 orderings of same-day picks.

• Drawdown is about halved. The worst single position cost about 1.5–2% of the account, against about 4–5% with flat 5% sizing.
• The idle-cash rule matters. Holding SPY only while the Balancer strategy is long SPY, and T-bills otherwise, is what keeps the drawdown down; holding SPY all the time does not.
• It gives up return in strong markets. The edge over SPY came from 2020-2023, which held two bear markets. In 2024-2026 holding SPY (+19.5% a year) beat every active version.

The Daily Picks

The daily picks publish every weekday before the US open.

• A standing list. A name that meets the entry rule keeps meeting it, and is republished every session it does. Each pick shows whether it is new and how many sessions in a row it has qualified. Names that have been on the list a while did not do worse than fresh ones.
• Sorted by distance below the 50-day average. Deepest first. Picks flagged ⚠ PRICE GAP (a 40%+ one-day move in the last 60 sessions, usually a split, spin-off or data error) go last.
• Stop and size on every pick, from the rules in How to Trade It.
• Buying conditions. A day-level reading — Weak, Normal or Strong — from the average setup score of the day's new picks. The score separates days, not the picks within a day. On the holdout, trades opened on weak days averaged about +1.8%, on normal days +5.7% and on strong days +6.1%.
• Idle cash. Whether idle cash should be in SPY or SGOV that day.

Data Quality

The OVTLYR feed occasionally has bad-tick errors — a single-day price that's disconnected from the days around it and disagrees with other public sources. Without correction these produce fictional trades in the backtest. The pipeline scans every ticker at load time for spike-then-revert patterns, statistical outliers, and OHLC inconsistencies, and repairs high-confidence cases in memory. The effect on the headline numbers is under 0.1 points of win rate.

Caveats and known limitations

Tools

The production scripts share the same entry and exit logic, universe filters, rolling classifier and risk rules. The live tools and the backtest make the same decision for any ticker and date.

---

Keep reading