Intermediate-term allocation with drawdown filters

A mix that can sit anywhere

A working allocation design sits between a nearly static mix and daily all-or-none switches. It may be fully invested, fully defensive, or anywhere between. The intended job is tactical asset allocation on an intermediate-term horizon: catch major uptrends and avoid major downtrends, and change the mix only when an intermediate market swing is identified.

The rotation cadence is bounded. The model is designed not to generate more than three or four round trips a year at the maximum, and typical years produce fewer, because higher signal frequency increases costly whipsaws. Average time in or out of a position is described as six months to a year, not daily or multi-day flips, so entry, exit, and abstention can be tested as one intermediate-term procedure.

Rank what remains after the loss bound

Rank-rotation is a procedure that orders candidate funds or exposures by a risk-adjusted score and then rotates allocation toward those that pass the filter, instead of picking names from last period's league table.

Candidate funds are ranked by risk-adjusted return after a program drops any name whose historical drawdown exceeds the maximum loss the account owner has said they can stomach. The drawdown-limit is a pre-trade and in-position bound that excludes any candidate whose historical peak-to-trough loss exceeds the loss the account owner has said they can live with.

A 10 percent historical hit is used as an example of a comfort bound. A fund that looks like a home-run candidate is still excluded if the owner has only been able to live with that smaller loss.

Judge the procedure over a market cycle

Absolute return is treated as an incomplete score. A result that looks decent in raw units can fail once the risk taken to earn it is included, so comparison needs a risk-adjusted baseline such as a Sharpe-style reading. The Sharpe ratio asks whether an absolute return is still attractive after the path of losses required to earn it.

Evaluation is not complete until at least one full market cycle has elapsed, described as four or five years or more, rather than a single rating-year sample. A market cycle is a multi-year stretch long enough to include both an advance and a decline.

Chasing last year's winner is rejected as a ranking rule because last year's winner is seldom next year's winner. The intended output is a durable procedure, not a one-year league-table rotation.

Allocate to the market first

The intermediate-term model is not purely technical. It uses three indicator categories, including a lower-weighted fundamental set that would have little value in a short-term system.

Market direction is treated as the dominant driver of individual stock moves, with an estimated 60 to 70 percent of up or down action attributed to the overall market rather than company-specific prospects. That is why allocation to the market itself is the primary construction decision.