Every other strategy here works from price history. This one works from the order book — the resting orders that have not traded yet — which makes it the only strategy in the collection looking at intent rather than at outcome.

It also carries the most interesting piece of risk machinery in the system: a gate that refuses the trade when the measured edge is below the fee floor.

Reading the Book Instead of the Tape

An order book holds resting limit orders on both sides — bids below the current price, asks above. The quantities are visible, and the comparison between them is the strategy's entire input: substantially more size resting on the bid side than the ask side is a bid-heavy book.

That is different in kind from every price-based measure. Price records completed transactions; the book records unexecuted intentions. The claim is that when intentions are lopsided, price tends to move toward the thinner side, because there is less resting there to absorb an advance.

The strategy enters in the direction of the dominant side and exits as the imbalance neutralises. The exit condition is symmetric with the entry: the position exists while the imbalance does, and the disappearance of the reason is the reason to leave.

Position size comes from the platform's edge model rather than being fixed — the strategy commits in proportion to how large the measured imbalance is, so a marginal reading produces a small position and a pronounced one a larger one.

When Book Pressure Leads Price

The condition is stated precisely in the source material: it works when book pressure leads price. That is an honest way to put it, because it makes clear the relationship does not always hold. Sometimes the book leads, sometimes it follows, and sometimes it is unrelated.

It needs a book with real depth. A thin book shows dramatic imbalances from small orders, and the ratio between two small numbers is not information. Deep, liquid pairs give the measure something to measure.

It also needs the measurement to be current, which is the hardest requirement here. A book changes continuously — orders are placed and pulled constantly — so an imbalance observed a moment ago may not exist now. This is the only strategy in the set whose input can be stale in seconds.

The Edge Gate

Entry generally crosses, since acting on a current imbalance means acting now, so the taker fee applies. Exit on neutralisation usually crosses too.

What distinguishes this strategy is the gate in front of it. The risk layer estimates the expected edge from the imbalance and compares it against the round-trip cost, and when the estimated edge is below that floor the trade is throttled — the strategy is prevented from taking it.

This inverts the usual arrangement. Most strategies discover their costs after the fact, in aggregate; this one is refused entry in advance when the arithmetic does not work. A signal being real is not sufficient — it has to be large enough to pay for acting on it.

The practical effect is that a strategy which could fire constantly fires selectively, and the selection is made on cost rather than on conviction. Most detectable imbalances are small, most small edges do not clear a round-trip, and so most valid signals are correctly declined.

The Book Can Lie

The central problem is that resting orders are not commitments. An order can be cancelled at any moment before it trades, and placing orders with the intention of cancelling them is a known behaviour — visible size that evaporates on approach. A strategy reading the book as intent is reading something that can be constructed to be read.

So the input is uniquely manipulable. Price history records what happened and cannot be retroactively altered; the book records what is claimed and can be withdrawn. Of all the inputs used across this collection, this is the only one that can be shown to the strategy deliberately.

The second failure is staleness. By the time an imbalance is measured, transmitted and acted on, the book has changed. The strategy is always trading a slightly historical picture of a structure whose whole value is being current.

The third is that imbalance neutralisation — the exit condition — happens in two very different ways. The thin side can fill in as new orders arrive, or the heavy side can be pulled. Both neutralise the reading and mean opposite things about what just happened, and the strategy exits identically on either.

The gate creates its own difficulty too, though it is the right kind. Throttling below-cost trades means the strategy sits out most of what it detects, and a strategy that mostly declines is hard to evaluate and easy to suspect of being broken. Its correct behaviour and its failure look similar from outside.

Order flow imbalance is the only strategy here reading intent rather than outcome, which is both its distinction and its weakness — intent can be withdrawn. The edge gate in front of it is the more transferable idea: a signal has to be worth more than it costs to act on, checked before the trade rather than discovered after it.

Note: This explains how a process works. It is not legal advice, it is not specific to any debt, and it is not a substitute for a licensed attorney in your state. Rules and time limits vary by state and change over time — check the cited sources.