Every performance report I get sent uses the same yardstick. Someone is up some number for the year and they hold it against a 60/40, or against just buying Bitcoin and sitting on their hands. And if they beat it, the conclusion is that the strategy works. If they lagged, the market was hard. Both conclusions are usually wrong, because the benchmark they picked cannot answer the question they actually care about.
The question is not whether you beat a fixed index. It is whether you captured the strongest moves that were sitting right in front of you. A portfolio that rotates between assets, sectors, or venues is implicitly claiming that it can find where the strength is and get there in time. A fixed benchmark never tests that claim. It just tells you whether you were long the right static thing.
Why a static benchmark hides the interesting failure
Say you run a rotating book and you beat a 60/40 by a comfortable margin over a year. Feels good. But if that year had a handful of assets that ran hard while the index dawdled, the honest comparison is not against the index. It is against how much of that available strength you actually grabbed. You can beat a lazy benchmark by a mile and still have left most of the real move on the table.
The reverse trap is worse. In a year where almost everything drifts sideways, a rotating strategy churns through trades, pays spread and fees on every switch, and ends up roughly flat. Held against a flat index, it looks fine. Held against the best you could have done by picking the two or three things that actually moved, it looks like you paid a lot of costs to accomplish nothing. The static benchmark lets that failure hide in plain sight.
What you want is a benchmark that moves the way opportunity moves. Something that says, for each period, here is what the leaders were doing, and here is how close you got to riding them.
Building a rotating leader benchmark
The idea is simple even if the bookkeeping takes a minute. Instead of one fixed asset, you construct a synthetic portfolio that holds whatever was leading on relative strength each period, and you let it re-select on the same cadence you actually trade. Then your real returns get measured against that.
A workable version looks like this:
- Pick a universe. Not the whole market, just the pool you would realistically rotate through. Maybe a dozen assets or sectors you actually consider. A benchmark built from things you would never touch is not a fair mirror.
- Pick a ranking period and a rebalance period. A common setup ranks by trailing relative strength over the last one to three months and rebalances monthly. Match the rebalance cadence to how often you really move, otherwise you are comparing yourself to a strategy that trades at a different speed.
- Each rebalance, rank the universe by return over the ranking window and take the top slice. Top two or three names, equal weighted, is plenty. This is your leader basket for the next period.
- Hold that basket, collect the return, then re-rank at the next rebalance and rotate. Chain those period returns together and you have an equity curve for a portfolio that was always sitting in the recent leaders.
That curve is your rotating leader benchmark. It is not investable in a clean way, because it uses last period's ranking to hold this period, and by the time you know the ranking the move has partly happened. That lag is the point. It represents a reachable-ish upper bound, a stand-in for what a decent momentum rotation would have captured. You are not trying to match it exactly. You are trying to see the gap.
Reading the gap
Once you have both curves, your real portfolio and the leader benchmark, the difference over time tells you something a static index never could.
If your returns track close to the leader benchmark, your rotation is doing its job. You are finding strength and getting there roughly on time. If you are consistently well below it while still making trades, that is the diagnosis you actually needed. Your rotation is adding activity without adding capture. You are trading, paying for it, and landing somewhere short of what a mechanical top-strength rotation would have done for free.
There is a third case worth naming. Sometimes you beat the leader benchmark. That usually means one of two things. Either you sized into a leader harder than an equal-weight basket would have, which is genuine skill in position sizing, or you got lucky with concentration and the same behavior will hurt you in a different period. Worth pulling those episodes apart and looking at them one by one rather than banking the outperformance as proof of anything.
A few things to watch so the comparison stays honest:
- Keep the benchmark's cadence matched to yours. If you rebalance weekly but benchmark against a monthly rotation, you are measuring against a slower animal and the gap is meaningless.
- Charge the benchmark realistic costs. A top-strength rotation that switches names every period pays spread too. If you compare your after-cost returns to a frictionless benchmark, the benchmark will always look unfairly good.
- Do not over-tune the ranking window until the benchmark flatters you. The point is a rough honest mirror, not a curve you optimized to feel better about your fills.
- Look at the gap across regimes separately. A strategy that keeps pace with leaders in trends but falls apart in chop has a specific, fixable problem, and one blended number hides it.
The main thing this changes is what a good year means. Beating a fixed index stops being the bar. The bar becomes how much of the available cross-asset strength your rotation actually converted, period by period, after costs. Most of the time the answer is humbler than the raw return suggests, and that is exactly the feedback that makes the next quarter's rotation a little less busy and a little more deliberate.