How a divergence model fought TSLA's moves both ways and edged ahead of the stock

TSLA was a volatile two-way mover over the window, the kind of tape a divergence rule is built for, and the question is whether reading those splits actually paid: here it traded both sides evenly and finished a step ahead of the stock with an unusually contained drawdown. This is a backtest of one fixed rule set replayed through TSLA's saved history, not a forecast or a recommendation, and the live sample is still too small to grade.

Disagreement is the quarry here. The model tracks TSLA's price beside a monitored reading and watches for the two to diverge, price driving one way while the reading tips the other, treating that split as evidence the prevailing move is tiring. New price ground the reading refuses to ratify triggers a sell, the inverse triggers a buy, and that even-handedness is why the ledger finished perfectly balanced, thirty-four longs against thirty-four shorts. Of the 68 trades, 57.35% were winners, accruing +180.53% at a flat stake.

How the model is built, end to end.

The reading the model lives by is the relationship between TSLA's price and a monitored indicator, not either one alone. When price and the reading move together, the model sees a move that still has support and stays out. When they split, price stretching to a new extreme the reading declines to follow, the model treats the gap as evidence the move is weakening. The size and direction of that divergence is the whole signal, which is what lets the rule lean against the prevailing move rather than with it.

A divergence is the trigger, not the trade. When price and the reading pull far enough apart the model opens against the tiring move, short into an unconfirmed rally or long into an unconfirmed slide, with a stop and target set in advance. From there the trade runs on rails. Because it is fading a move rather than chasing it, its stops tend to be close, which on TSLA produced a near-even split between targets and stop-outs, with a smaller group timing out when the expected turn neither arrived nor clearly failed.

The illustration above walks the model across TSLA in a single line: the price track, the monitored reading beside it, the divergence between them, and the long or short that divergence produces. Nothing in the chain is elaborate; the character is contrarian patience, waiting for price and the reading to disagree before leaning against the move. The model's lineage belongs in the frame too. It began as one candidate in an automated search and reached deployment only by clearing backtest and walk-forward checks, the route from idea to live rules behind every figure here.

At a glance

These figures come from a backtest of the model on TSLA, scored against fixed acceptance gates, not from a live track record.

Walk-forward verification

A trade walked through

The walked example is a long held about three days. With TSLA falling while the monitored reading refused to confirm the drop, the model bought near 383 dollars and the rebound carried it to its target near 423 dollars for +10.38%. It is the model working as designed on the long side, fading a tiring decline and exiting at the target once price turned.

What the full trade record shows

Across its 68 TSLA trades the model won 39 and lost 29. The exits split closely between the stop and the target: 29 were stopped out, 27 reached the profit target, and 12 closed on a time exit.

A near-even stop-and-target split is the signature of a contrarian rule with close stops: fading a move means many trades are knocked out when the move persists, while the turns that do arrive are taken cleanly at the target, with a smaller share timing out.

The biggest winner was a short into an unconfirmed rally that reached its target for +11.76% over about eight days in July 2024. The book's winners came in both directions, reflecting a model that took long and short in equal numbers as divergences pointed each way.

The worst trade was a short faded into a rally that kept climbing, stopped out for -4.32% over roughly two days in December 2024; that single loss is also the equity curve's deepest drawdown, which the close stops held to under five percent.

No single trade carries the record on a book this size, and the lesson of the set is that fading tiring moves on a volatile two-way stock kept losses small and edged the rule past the stock, even with an ordinary win rate.

How does this compare to just holding TSLA

Over the same window the model was tested on, buying TSLA and holding would have returned a little less than the rule did, so here the comparison favors the model. Setting the two together shows it earned its keep on this tape rather than merely riding it, and the tiles below put numbers on the edge.

How well does the model reproduce its tape?

Walk-forward verification checks whether the saved rule path reproduces the expected signal behavior on held-out timestamps it was not built on. It is a consistency and replay-integrity test, not proof of live profit. A clean reproduction means the deployed rules behave like the studied ones; it says nothing about whether TSLA's price and the monitored reading will keep diverging the way they did across the test window.

In live trading the model has been quiet so far, with too few signals to set beside the backtest. Until more live trades accumulate, the backtest is the only evidence on hand, and it should be read as exactly that, a study of how the rules behaved on saved history.

When this approach fails

The model's failures come from the same place as its trades, the divergence. Its standing risk is fading a move that simply keeps going: price and the reading split, the model leans against the move, and the move extends anyway, which is how the deepest trade lost 4.32% on a short into a rally that refused to roll over in December 2024. Tight stops keep any single loss small, but a sub-threshold win rate and a thin average return per trade are what left two of the five acceptance gates unmet.

Three things are worth watching if this ever trades at size. The first is the contrarian risk, since fading a move that keeps running is how a divergence rule bleeds. The second is the slim average return per trade, which alongside the win rate is what left the expected-return gate unmet despite a tight drawdown. The third is the gap between live and backtested behavior, the first sign price and the monitored reading are no longer diverging the way they did in the study.

Risk and honest limits

On this run the model's automated checks logged a caution rather than a clean pass. It is one more reason to read everything here as backtest evidence about the rules, not a verdict on the stock.

Lifecycle