SystemTrader's Sirius is a trend-quality strategy. It asks not only whether a stock has risen, but whether it rose well — steadily, on consistent daily strength, repeatedly making new highs along the way.
This page explains the three measures Sirius combines, why a stock must rank well on all of them rather than excel at one, how the tradable universe is chosen by rank rather than by any fixed threshold, and how positions are sized and protected.
Two stocks can post the same six-month return by completely different routes. One gapped 40% on a single announcement and drifted sideways for months. The other ground higher week after week, closing at new highs dozens of times. A plain return ranking cannot tell them apart. Sirius can, and it prefers the second.
It scores each candidate on three separate measures of the same six-month window and blends them equally. Because the blend is equal, a stock that is exceptional on one measure but poor on the others does not rank — consistency across all three is the requirement.
The plain six-month price return. On its own this is the measure most easily faked by a single gap, which is exactly why it is only one of three.
The balance of up days against down days, measured from each session's own open to its close rather than from yesterday's. That distinction matters: it captures whether buyers controlled the day itself, not whether the stock happened to open higher on overnight news.
The share of sessions that closed at a new six-month high. A stock making highs repeatedly is in a trend that keeps being confirmed; one that made a single high months ago and has hung below it since is not, even if the return still looks similar.
Each measure is converted into a percentile against every other eligible stock that day, and the three percentiles are averaged. Nothing is compared to a fixed threshold, because what counts as a strong six-month return in a calm year and in a violent one are not the same number.
Holdings are not simply the top-ranked names. A stock already held keeps its place while it remains within a wider rank band, so the portfolio is not churned every time two names swap positions. Ranks move constantly; trading on every reshuffle costs more than it earns.
Sirius trades the most heavily traded common stocks, selected by rank. This is a deliberate design constraint rather than a detail. A fixed dollar or share threshold means something different in every market: the same cut-off can admit a few hundred names in one period and twice as many in another, quietly changing the strategy without anyone choosing to.
Ranking keeps the universe the same size in every environment, so the historical record reflects the rules rather than the drift of a number picked years ago.
A name is eligible to be bought only while it trades above its own short-term moving average. A stock can still rank well on six months of history while currently rolling over, and this condition keeps the strategy from buying into that decline.
It applies to purchases only. A position already held is never force-sold because it slipped below the average — it leaves on rank alone. Being harder to enter than to hold avoids selling a good position into a brief dip and immediately rebuying it.
Positions are not equal-weighted. A calmer stock receives a larger allocation and a more volatile one less, so each contributes a more comparable amount of risk rather than a comparable amount of money. A cap prevents any single quiet name from dominating the book.
Positions are then left to drift and only resized once they move well away from target, so ordinary daily movement does not generate trades.
Overall exposure is reduced when short-run market volatility rises above its own longer-run level, and restored as conditions settle. Separately, while the broad market trades below its long-term trend, exposure is halved again and no new positions are opened.
Exits continue to work normally in those periods — the strategy will still sell a position that loses its rank. It simply stops adding risk while the market is falling.
Scores are computed after the close from that day's completed data, and trades execute at the next morning's open. Signals publish in the evening, so you have the full night to place orders. No decision uses a price that had not yet printed when the decision was made.
All performance figures are based on historical backtesting and are hypothetical. Past performance does not guarantee future results. Sirius is a long-only momentum strategy and falls with the broad market; its worst historical stretch was a drawdown of roughly a quarter of the portfolio. The exposure overlay reduces risk in falling markets but does not eliminate it, and it can also keep the strategy underinvested through the early part of a recovery. The backtest assumes execution at open prices with no commission, spread, or slippage; real-world trading will differ. This is educational content, not investment advice.