Why a trend-vote model turned a profit while COIN went nowhere

COIN spent the window churning rather than trending, finishing below where it began, and the story worth telling is how a busy, two-sided model squeezed a gain out of a tape that left buy-and-hold underwater. What follows runs the model's fixed rules forward over COIN's saved history as a study of the rules, not a forecast or a recommendation; the live sample is still too small to grade.

On COIN's falling, jittery tape the design found its element. Several momentum gauges, running from short horizons through long ones, each cast a vote on direction, and the model commits only once a clear majority points the same way: a ballot tilted up turns into a buy, one tilted down into a sell. Because the shares spent so much of the window sliding, the sell side dominated, accounting for 115 of the 198 trades. That heavy churn won 45.96% of the time and, carried by a handful of clean breakdowns, totalled +337.89% on a fixed stake.

How the model is built, end to end.

The reading the model lives by is directional agreement across COIN's trend clocks. When the short, medium, and longer look-backs point every which way, the model treats the tape as noise and waits; when most of them swing the same direction, it reads that as a trend worth joining. The reads are scaled to COIN's own swings, so what counts as agreement is set by how this stock normally moves. A scattered set of look-backs keeps it flat; a lopsided one opens the door to a trade either way.

A majority vote is permission, not a trigger. The model still waits on a separate direction read to agree before it acts, which keeps it out of moves that line up for a moment and fall apart. When the vote and the confirmation match it places a position with a stop and a target fixed up front. From there the trade closes in one of three ways, and on COIN the protective stop did the most work, shutting more positions than the target or the settling reading.

The illustration above carries the model across COIN start to finish: the several trend reads, the majority they must reach, the direction check, and the long or short that follows. Nothing in the chain is fancy; the model's character is in demanding that the look-backs agree and the read confirm before it puts money down. Its lineage sits in the picture as well, it surfaced as one option among many in an automated search and stuck around only by clearing backtest and walk-forward checks, the route from idea to deployment behind every number on the page.

At a glance

These figures come from a backtest of the model on COIN, 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 just under a day. COIN's trend reads swung up together near 191 dollars, the direction check confirmed, and the model bought; the move ran quickly and the trade closed near 221 dollars for +15.92%. It is the model catching a clean, short burst, in for the agreed direction and out before the alignment could fray.

What the full trade record shows

Across its 198 COIN trades the model won 91 and lost 107. The exits leaned on the stop: 98 were stopped out, 61 reached the profit target, 38 closed when agreement settled back, and one closed on a late signal.

A stop-heavy split like this fits a busy two-sided model on a choppy stock: nearly half the positions are knocked out when an agreed direction fails to follow through, while the winners that survive are taken either at the target or as the vote settles. The book leans on its target hits to outweigh the frequent stops.

The biggest winner was a long that ran for +38.05% over about seven days in June 2025, entered near 253 dollars and closed near 349 as the settling reading let it run past a hard target. The model's best trades came when an agreed direction carried cleanly before the look-backs fell out of step.

The worst trade was a short stopped out for -8.98% over about six days in early March 2025, entered near 213 dollars; that loss marks the equity curve's deepest drawdown of 8.99% and is why the drawdown gate is one of the three that bites.

No single trade carries the record across 198 of them, and with COIN finishing the window lower, the lesson of the book is that the model's edge here was its freedom to short, turning a falling tape into a gain a passive hold could not.

How does this compare to just holding COIN

Over the same window the model was tested on, simply buying COIN and holding it would have lost money. Setting the two side by side shows whether the rule earned its keep, and here it plainly did, turning a profit on a stock that finished lower, mostly by being free to short. The tiles below put numbers on the gap.

How well does the model reproduce its tape?

Walk-forward verification checks whether the saved rule path reproduces the expected signal behaviour on held-out timestamps it was not built on. It is a consistency and replay-integrity test, not proof the model will make money live. A clean reproduction means the deployed rules behave like the studied rules; it says nothing about whether COIN will keep chopping the way it did across the test window.

In live trading the model has been quiet so far, with too few signals to line up against the backtest. Until more live trades accumulate, the backtest is the only evidence there is, 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 losses come from the demand at its core, agreement that does not hold. Its deepest single trade was a short near 213 dollars that COIN rallied against, stopping out for -8.98% in early March 2025, the same stretch behind the 8.99% drawdown. With a 45.96% win rate it loses more trades than it wins outright and leans on a few big directional moves to carry the book, so a run of votes that reverse soon after entry is its main hazard, and the heavy trade count means small losses can pile up quickly.

Three things are worth watching if this ever trades at size. The first is the sub-half win rate, because the record depends on a small number of large winners outpacing many small losses. The second is the drawdown near 8.99%, which together with the high stop rate means losses can cluster. The third is the distance between live and backtested behaviour, the first sign the agreement reads are no longer marking the same moves.

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