Testable cycle rules as one procedure

One checklist, three allowed outcomes

A documented cycle-and-intermarket workflow is reviewed here as a single mechanical-trading-system. Rule inputs, market state, and execution-constraint limits must feed one written procedure. That procedure is allowed to emit an entry, an exit, or an abstention-rule over the system's holding period.

TradersWeek editorial reading: the archive steps are not a menu of optional charts. They are one testable sequence. A signal is not treated as live until a cross-platform-audit rebuilds the same rules and data on a second platform.

Date the cycle before a stock pick

The first step in the documented procedure is a global-market and business-cycle read that dates expansion or contraction before any single-country stock pick is considered.

That read is the business-cycle-template. Since the 1960s the documented procedure uses an average bull-market length of about 4.3 years only as a template for early, mid, or late positioning, not as a forecast of the next top.

Phase filters and allowed leadership

Early-bull confirmation in the procedure requires consumer discretionaries to outweigh staples on a ratio and treats financials as an early leader in both directions. That comparison is a sector-leadership-ratio: a confirmation input, not a standalone forecast.

Under the declining-dollar overlay used from 2002 onward, dollar strength is treated as a defensive warning and euro strength versus the dollar as a risk-on green light. That check is the currency-regime-filter. It frames risk appetite before equity or sector signals are allowed to fire.

The compiled intermarket map sequences sector rotation, five-year yields, 30-year bond prices, the yield curve, GDP, and unemployment across late contraction through early contraction so each phase has a different allowed leadership set.

Confirm a move without assuming a cycle

A moving average is reserved for longer trends, while shorter slope checks are done with a straightedge against the screen, and both exist only to confirm that a trend is still in force. Those checks are the trend-confirmation-tool set.

Recognizable clustered lows may justify watching a time window for a probable trough, but the same rule set forbids assuming a cycle will appear, and shallow or missing lows remain valid outcomes. That band is a cycle-low-window. When the window, the intermarket map, or an execution-constraint fails to line up, the written procedure must return an abstention-rule.

Factory output still needs an audit

A strategy-builder can emit on the order of 100 candidate rule sets from one year of 30-minute exported data, including EasyLanguage code, after a roughly 10-minute setup. That pile is strategy-factory-output. Editorial reading: generated code is a candidate list, not a live procedure.

Sorting factory output by percent wins can surface a 100% win set with only four trades and an average trade of $10.25, which the review treats as too thin to size or compound. Sample size, average trade size, and compounding capacity are execution-constraint limits that can veto an otherwise positive backtest.

The same pasted rules produced conflicting audits: about $1.96 average trade in the builder versus $41.11 on the originating platform, with more trades and a slightly lower win rate on the second run, so a positive net-profit print is not accepted as a finished system. That second-system rebuild is the cross-platform-audit.

What the procedure is for

Editorial close: the archive workflow is useful as a decision process when every filter is written down, every phase has an allowed leadership set, and a sit-out outcome is a valid result. The holding period of the system, not a forecast of the next top, is the horizon that those rules are meant to serve.