The Museum Exhibit Resonanz Capital — Spotlight Strategy Notes By Vincent & Saâd
A manager sits across from you and hands you a ten-year track record. Clean. Sharp. Low drawdown, no bad years. The dates: 2010 to 2019.
What have they actually given you?
A record of decisions made in a world that no longer exists. Not slightly different — fundamentally different. And too many allocators are still treating that exhibit as a roadmap.
This piece works through the argument: when pre-2020 history misleads, when it still travels, and what a more useful evaluation framework actually looks like.
The core problem: non-stationary data
When the environment shifts structurally, data from the old environment doesn't just become less relevant — it can actively mislead you. You are reading numbers from a world that doesn't exist anymore.
The structural breaks are not subtle. Rates went from falling for twenty years to rising sharply. Inflation went from effectively zero to persistent. The correlation between equities and bonds — the foundation of multi-asset portfolio construction for two decades — flipped. These aren't temporary disruptions. They are changes to the whole setup.
The instinct to value a longer track record makes sense. A manager who has been running for ten years has been through things a two-year manager hasn't. The problem is that logic relies on an assumption nobody actually said out loud: that the environment those decisions were made in is still similar enough to today. For a lot of strategies, it simply isn't.
The direction of the error matters
The argument above can be oversold. The claim cannot be "pre-2020 history is unreliable, full stop." The direction of the mistake matters.
Consider managed futures and trend-following CTAs. Those strategies had a genuinely tough 2011–2019: low volatility, choppy markets, persistent drawdowns. If you had discounted their pre-2020 history and concluded the strategy was broken, you would have missed 2022 entirely — their best year in a generation. The pre-2020 record actually understated what those strategies could do.
The correct framing: pre-2020 history is unreliable in specific ways, depending on how sensitive that strategy is to the old environment. That becomes the first question in any manager evaluation.
Strategies with high regime sensitivity — long-duration fixed income, correlation-dependent multi-asset construction — deserve heavy discounting on their pre-2020 return numbers. Strategies with low structural bias and high decision frequency travel better. But even then, run the attribution. Know what you're actually looking at.
A note on attribution: it's messier than it sounds. The models you use are themselves sensitive to the environment. You can run the same track record through two different models and get genuinely different answers about whether a manager has skill. Treat attribution as a starting point for judgment, not a final answer.
Count bets, not months
There is a separate dimension most allocators miss, and it is not about time. It is about decisions.
A high-frequency strategy making hundreds of independent decisions a month — six months of that gives you real information. A concentrated long-only manager holding twelve positions for years at a time might have a five-year track record and have made very few truly independent calls. You cannot compare those on length alone.
This bites hardest in exactly the strategies most affected by the regime question. The strategies most contaminated by pre-2020 conditions are often also slow-moving structural bets — on rates, credit spreads, equity risk premia. Thin decision count and a contaminated sample. Both problems at once.
2020 and 2022 are not the same test
These two years get conflated constantly. They shouldn't be.
2020 was a liquidity crisis — sudden, violent, and then over quickly. Central banks came in fast. What it tested: can you survive a severe short-term shock without blowing up? Most managers did survive 2020, which is why it's now a weak differentiator. Everyone sitting across from you today has a "navigated COVID well" story.
2022 was the opposite: slow, grinding, a full year of losses. Equities down roughly 20%, bonds down 15–20% at the same time. No hedge worked, and no rescue came. The central banks weren't coming to help. You had to hold your positions based on your actual conviction.
That is why 2022 is the more useful stress test for today — not because it was worse in the moment, but because its structure is more representative of the current environment: inflation, central banks tightening, the bond-equity hedge not working.
A manager who navigated 2022 well, and you can understand why, tells you more than any three years from 2010 to 2019.
That said, making this a hard rule creates its own selection bias. Some very good managers launched after 2022 — former senior PMs from strong platforms, long experience, strategies they've run before. Apply "must have been live in 2022" as a threshold and you exclude them entirely, not because of any real signal about their quality, but because of a calendar. And 2022 was one specific type of stress: sustained rate rises. It doesn't test credit crises, deflation, or currency crises. Making it your main filter doesn't remove recency bias. It just swaps one for another.
For managers not live in 2022: go sideways rather than backwards. What did this person do at their prior firm during that period? How did they communicate through it? Use whatever track record exists to examine process consistency rather than return numbers.
The consistency test
Some of the most convincing explanations of bad performance come from managers who are great at talking and genuinely poor at investing. Explaining things well after the fact is a skill — it is not the same skill as investing. A clear explanation of a tough year is necessary. It is not sufficient.
The real test is not whether they explained it well. It is whether the explanation is consistent with what they were saying before the tough period. If the story matches their prior letters, their risk framework, and how they were positioned going in, it means something. If it is a new story that conveniently fits a bad outcome they did not see coming, that is noise.
Pull letters from before the difficult period. Does the positioning they describe match the drawdown you see? Consistency over time is the signal, not the quality of the post-hoc narrative.
The hard case
Any framework has to work on difficult examples.
A manager with a fifteen-year track record, market neutral long/short equity, genuinely low net. Pre-2020 record looks strong. 2022 was bad — not a disaster, but a real drawdown in a year when a market neutral book should not have been badly hurt. Their explanation: value factor exposure caught in a crowded trade unwinding in Q2. They've adjusted risk monitoring since. Team is stable, process is well documented.
The framework doesn't give you a clean answer here. The 2022 performance looks like a warning sign — the strategy did the thing it was not supposed to do. But the explanation is plausible. Market neutral funds with value exposure genuinely did get hurt by factor crowding in Q2 2022. That was partly a market structure problem, not purely a process failure. And the pre-2020 record, even if you discount the return numbers, shows a team that ran risk consistently for fifteen years without a serious blow-up.
The right move: sit with the uncertainty. Put more weight on the next twelve months of live performance than you normally would. Talk to the prime broker. Look at the timing of the drawdown — does it actually match the factor crowding story? Do the work. But you do not get a clean answer, and pretending you do would be wrong.
A framework helps you think. It does not replace your judgment.
What this means in practice
How deep you go depends on who you are. A large institution with a full research team can run attribution, build a proper factor model, and construct a comparison set. A smaller family office is making the same judgment with two people and limited tools. Same principles, different depths.
In both cases:
Start with regime sensitivity. Is this strategy's performance driven mainly by conditions that existed from 2010 to 2019 — falling rates, low volatility, the equity-bond hedge? If yes, discount the return numbers heavily. Do not discard the process evidence from that period. Just the numbers.
For strategies independent of that macro setup — high decision frequency, low structural bias — the history travels better. Still run the attribution.
Weight 2022 heavily as a stress test, but not as a hard rule. Use it as the best available test for the current environment while knowing it was one specific type of stress. For managers not live in 2022, go sideways: prior firm history, how that team behaved during the period, comparable funds.
Apply the consistency test carefully. What matters is whether the story holds up over time, not whether it sounds good today.
Count bets, not months. For concentrated strategies with slow decision cycles, adjust your confidence in the track record accordingly.
Track records are evidence — but evidence of something more specific than most evaluation processes acknowledge. The job is knowing which question each part of the record is actually answering.
The museum exhibit is the right frame for the return numbers produced in a fundamentally different environment. It is less apt for the process evidence underneath. Know which part of the exhibit you are looking at.
