The four-guardian framework is editorial language for the engines that drive the Shishin system. It is not a separate product, a regulatory entity, or a claim about market forecasting. The engines and their behaviour are described elsewhere in the methodology library; this piece is about why we wrap them in this particular set of names.
Four engines could have been called Engine 1, Engine 2, Engine 3, Engine 4. Most quant teams stop there. We deliberately did not. The four-guardian wrapper does something the codebase cannot do on its own, it makes the regime call legible to the operator at a glance, before any numbers are read.
The four symbols
The Four Symbols of East Asian cosmology, the four constellations that historically governed the directions of the compass and the seasons of the year, map unusually cleanly onto a four-engine regime-switching architecture. Each carries a temperament. Each implies a posture. The mapping is below.
- Genbu · 玄武 · the black tortoise of the North, the symbol of endurance and slow accumulation. Maps to the quality-composite engine. Fires when the regime favours patient, durable margins and conservative leverage. When Genbu is awake, the system is not chasing; it is accreting.
- Suzaku · 朱雀 · the vermilion phoenix of the South, heat, ascent, momentum. Maps to the small-cap momentum engine. Fires when breadth is expanding and small-cap breakouts are working across the market. Suzaku is the attacking posture.
- Byakko · 白虎 · the white tiger of the West, defence, ferocity in retreat. Maps to the defensive-sector engine. Fires when the broad market is hostile to small-cap momentum, but a narrow subset of names continues to work. Byakko is the defensive posture, not flat, but selective.
- Seiryū · 青龍 · the azure dragon of the East, speculation, recovery, large-cap surge. Maps to the recovery-large-cap engine. Fires during macro transitions out of weakness, when the early signal is a small number of mega-caps pulling the indices off the lows. Seiryū is the transition posture, one foot still in defence, one already moving.
What the framework actually does
The naming is editorial; the math is unaffected. The regime classifier decides which engine is active based on breadth signals computed across the full investable universe. It does not consult the cosmology. It does not care which direction Genbu defends.
What the framework does is give the operator a single word for the entire posture. When the system flips from Suzaku to Byakko, that reads, immediately, without opening the charts, as “we’re done attacking, time to defend.” A single line in the log conveys what would otherwise be a paragraph of breadth-state language. The framework is a compression layer between the math and the human.
Why mythology specifically
The choice of the Four Symbols over, say, factor labels (“momentum”, “quality”, “defensive”, “recovery”) is deliberate. Factor names are technically more transparent but emotionally inert. They are categorisations from academic finance and they carry the connotations of academic finance, precision, measurement, a sense of clinical detachment from any actual trading day.
Mythological names are the opposite. They carry temperament, season, time-of-day, posture. Genbu’s connotation of slow durable accumulation is not a metaphor for the quality engine; it is what the quality engine is actually for. Suzaku’s heat is not a metaphor for momentum; it is the operator’s correct mental state when the engine is firing. The framework gives the operator a shorthand for the posture the system is taking on any given day, and an emotional vocabulary for how to relate to it.
The discipline cost, calling our momentum engine Suzaku rather than “Engine B”, is nothing. The benefit, over years of running the system, is that the human in the chair never confuses Genbu season with Suzaku season. Naming structures attention. Attention structures behaviour.
What the framework does not do
It does not influence trade selection. Each engine has its own filter chain, its own composite score, its own sizing and stop policy. None of those reference the guardian name; the name appears in logs and dashboards, not in the trading code.
It does not predict regime transitions. The guardian-of-the-day is a reflection of the regime classifier’s call, not a cause of it. When breadth deteriorates we do not switch to Byakko because of any cosmological reading; the breadth signal triggers, the classifier flips, and the dashboard re-labels the active engine. The framework follows the math, never the other way round.
It does not turn the system into a Japanese-themed retail product. The mythology is editorial. The bot trades on numbers. The bot does not know it is called Shishin.
Why this matters editorially
A four-engine architecture without a unifying narrative is hard to live with. Operators end up referring to the engines by version number, factor description, or whichever ticker happens to be exemplary that week. The mental model drifts; institutional memory thins. The framework fixes the language. Five years from now, when the engines have been retuned twice and the parameters have moved, Genbu will still be what Genbu has always been: the patient one, the one we let do nothing for weeks at a time, the one that fires when the chase has gone out of the market.
That continuity, the guardian holds its meaning even when the implementation moves, is what a framework is supposed to do.