Risk On or Risk Off: The Six Factors Behind the Delta Engine

This factor looks at the shape of the weekly Bitcoin candle. The size of the candle body, the length of the upper and lower wicks, and how they relate to each other tell you whether buyers or sellers controlled the week.

A week that closes near its high with a small lower wick is a statement from buyers. A week that closes near its low with a large upper wick is a statement from sellers. Weeks that close in the middle with long wicks in both directions are undecided. The factor translates that visual story into a score.

Bullish candle shapes contribute toward bottom evidence. Bearish shapes contribute toward top evidence. Indecision candles contribute a smaller amount toward whichever direction the other factors are already pointing.

Price alone does not tell you who is in control. A price that rises on thin buying looks the same on a chart as one that rises on genuine demand. Delta volume cuts through that by measuring raw order flow: the difference between what was bought at the ask and what was sold at the bid across the full week.

When buyers are consistently outpacing sellers by a meaningful margin, it tends to precede sustained moves higher. When net selling accelerates to an unusual degree, it tends to precede pullbacks. The key word is unusual. Mild net selling is the normal baseline in this market. The factor only scores when the reading moves well outside that baseline in either direction.

High volume on its own means nothing. The same week of heavy trading can be bullish or bearish depending entirely on where price is when it happens. This factor scores volume only in context.

Heavy trading near a recent low suggests buyers are stepping in with conviction and sellers are exhausting themselves. That is a bullish read. Heavy trading near a recent high suggests the opposite: buyers are running out and sellers are distributing into the strength. Light trading near a high signals a move that lacks the participation needed to sustain itself. In the middle of a range, volume tells you little and scores accordingly.

The American Association of Individual Investors surveys its members every week and publishes the percentage who are bullish versus bearish. The Delta Engine uses this as a counter-indicator. When the crowd is most optimistic, markets tend to disappoint. When the crowd is most fearful, markets tend to recover.

The critical word is most. Mild optimism or mild pessimism tells you nothing useful. The factor only scores when sentiment reaches a historically extreme reading in either direction. Mid-range readings contribute zero, because mid-range sentiment has no reliable predictive record.

One practical note: the AAII survey is published on Wednesdays. The Monday signal is therefore preliminary on this factor. The confirmed reading updates on Thursday once the sentiment data has been manually entered.

This factor asks a simple question: is price near a structural extreme right now? Markets tend to find support and resistance at the edges of their recent ranges. A price sitting close to a multi-month low is in territory where buyers have historically stepped in. A price sitting close to a multi-month high is in territory where sellers have historically appeared.

The lookback window is 12 weeks. That is long enough to capture genuine structural extremes without triggering during normal short-term consolidations. Price in the middle of the range, where reversals are less predictable, contributes nothing to the score.

This factor watches for shifts in the direction of order flow from one week to the next. A week where buyers were in control following a week where sellers were in control is a meaningful change in character. The same logic applies in reverse. These direction shifts often appear before price itself reflects the change.

It is deliberately the smallest contributor in the framework. A single week of reversing order flow is an early signal, not a confirmation. Without the other factors reinforcing it, a direction shift alone is insufficient to drive a strong read. Its job is to add weight when the other factors are already leaning in a direction, not to drive the signal on its own.