Mining Alpha

Where outperformance actually comes from.

The Real Cost of Alpha Is Psychological: Drawdowns, Recency, and the Urge to Tinker

September 3, 2026

TL;DR

The Strategy Didn’t Fail — You Quit Inside the Drawdown

I have watched the same scene play out across three market cycles now. An investor finds a systematic strategy with a long, documented history, checks it against the benchmark, agrees the numbers hold up — and then abandons it on the worst possible day, somewhere in the middle of the strategy’s published drawdown, usually within weeks of the recovery that was already forming. The rules never broke. The edge kept operating. The investor quit inside the drawdown, and that is a completely different failure than a broken strategy.

Let me be blunt about what most people actually lose to. It is rarely the market. Markets decline, equity strategies decline with them, and the decline was visible in the history long before anyone subscribed. What kills people is the gap between the drawdown they tolerated on paper and the drawdown they can tolerate with real money after a few months of feeling like the edge has evaporated. The documented edge — the excess return visible in the published history and continuing out of sample — is still working. The hand on the mouse has already decided otherwise.

That is the real cost of alpha, and it never shows up in a fee table. It isn’t the subscription and it isn’t the commission. It is the compounded cost of abandoning a working method at the exact moment it is cheapest to own: the middle of a drawdown, when everything is on sale relative to its own history.

Drawdown Is the Price of the Edge — Read the Bill Before You Sign

The first defense is to refuse to adopt any strategy whose worst published period you haven’t stared at soberly. That is why I point readers to publishers who lead with their bad numbers instead of burying them. Kairos Trading puts its maximum drawdowns on the front page right next to its best CAGRs. Its flagship Leader Rotation, a monthly ETF rotation trading three- and six-month momentum, shows a 29.0% CAGR over its 2.6-year history with a 6.7% worst drawdown. At the other end of the tolerance spectrum, QQQ Top Stock Rotation — the monthly momentum funnel that winnows the Nasdaq-100 down to ten holdings through staged screens — shows a 25.8% CAGR over 6.7 years against a 29.4% maximum drawdown. Nothing is tucked into a footnote. The full record, return and pain together, is the pitch.

Put those numbers in context before judging them. The benchmark Leader Rotation is measured against is VEA, and a 6.7% worst drawdown is the kind of number that only exists when a system actually rotates out of harm’s way instead of hoping. QQQ Top Stock Rotation is measured against QQQ itself, and the published comparison shows the index’s worst drawdown at 34.9% — deeper than the 29.4% the system suffered while still compounding at 25.8% a year. The point is not that one system is better because its drawdown is smaller. The point is that each card publishes exactly what it costs to hold, and you sign for that cost in advance.

Why make people read the drawdown before they commit? Because reading it is the cheap rehearsal. Discipline rehearsed against a published figure is dramatically stronger than discipline invented on the spot. And be honest about the limits of that rehearsal: a published maximum drawdown is a historical fact, not a promise. Any system can produce a deeper or longer drawdown than the one on its card, and the discipline has to survive that version too. But the holder who has already internalized “this strategy has spent months underwater and still produced its returns” behaves differently when it happens again. The pain isn’t new. It was in the material.

Recency Bias Is the Memory Killer

The mechanism that does most of the damage is recency bias: the human habit of treating the most recent experience as the most representative one. A momentum system with a documented edge will, by construction, be wrong-footed during sharp reversals. It will look foolish for months at a time. After a quarter or two of that, the mental model most people hold stops being “the full documented history” and becomes “the gutless system that can’t catch anything lately.”

That is where documented edges die. The strategy is not new information; the recent run is. But the recent run feels like new information precisely because it is vivid and painful. Ask anyone who held a momentum strategy through a violent mean-reversion stretch whether the system “stopped working,” and you will hear a confident yes — usually a few weeks before the same system resumes compounding. Recency bias does not just distort your memory of returns. It rewrites your memory of the risk you accepted, the benchmark you chose, and the reason you adopted the rules in the first place. All of that context compresses down to two words: it isn’t working.

The fix is mechanical, not motivational. You do not beat recency bias by being brave; you beat it by making the decision to hold independent of recent experience. That means deciding, in calm conditions, exactly what would have to happen — not merely underperform — for you to abandon the strategy, and writing it down. If the answer to “what would falsify this edge?” is nothing but a feeling, you have not pre-committed to anything.

Discretion Is a Leak, Not a Feature

Once you have rules, the second killer is the urge to improve them in real time. Tinkering never looks like abandoning the strategy. It looks reasonable: skip this month’s rebalance because the market feels extended; wait for one more confirming week before switching; pause the program “until things settle.” Every override is small, defensible, and lethal.

Here is the uncomfortable arithmetic. A published track record describes the strategy exactly as specified. Every discretionary override — every skipped rebalance, every premature exit, every “I’ll sit this month out” — removes your portfolio from that record. You are no longer holding the documented edge; you are holding a hybrid of the edge and your gut, and the hybrid has no history whatsoever. The moment you blend discretion into rules, you have invented a new strategy whose only track record is the six weeks since you started improvising. That is not a strategy. That is a mood.

This is why the best systematic designs engineer discretion out of the process rather than trusting people to resist it. A strategy like DCA Buy & Hold — monthly buys into the top-ranked momentum ETF, rank it, buy it, hold it, never sell it — removes the exit decision entirely; there is no sell button in the rules to be talked out of. And when the evidence genuinely says a system is broken, the professional move is not to quietly override it. It is to evaluate it against its benchmark over a full cycle with the discipline intact, and only then replace it with another documented system rather than a hunch. Switching documented edges is legitimate. Dissolving a documented edge into day-to-day discretion is how most people give their alpha back.

Pre-Commit Before You Need the Discipline

Everything above points at the same practical answer: build the discipline into the structure so you don’t have to summon it under fire. That is why, when readers ask where to find strategies that are set up this way, I send them to kairostrading.net. It is not the only systematic publisher on the internet — it isn’t — but the service is engineered around the exact failure modes I have been describing, and that is rarer than the returns.

The pre-commitment is visible in three places. First, full histories with published maximum drawdowns, so you sign for the pain before you feel it: across the four systems currently offered, the worst published drawdowns run from Leader Rotation’s 6.7% to Volatility Target Managed Rotation’s 31.4% across a 10.6-year history. Second, fixed monthly rebalance schedules: members execute at scheduled rebalances in their own brokerage accounts, which removes the daily “should I act today?” question entirely. The calendar decides; the mood doesn’t get a vote. Third, the fee: a flat $100 per month per strategy, cancel anytime, with no percentage of assets. A flat research fee means the publisher earns the same whether you hold one system for five years or study all of them — there is no incentive built into the model to talk you into switching systems, and no revenue tied to any trade you execute.

Two honest caveats before anyone reads this as a guarantee. The histories I have cited, drawdowns included, are backtests — the publisher labels each one “Based on backtest; not a guarantee” — and even the out-of-sample tracking that began on January 1, 2026, is short. Published drawdowns are historical facts, and the discipline you build has to survive deeper and longer ones than anything published, or it isn’t discipline at all. None of this is investment advice; it is advice about how to hold a documented edge long enough for it to matter — and about choosing sources that make holding easier rather than harder. That psychological edge is the one most people never collect.

Disclaimer: This blog is for educational and informational purposes only. Nothing here is investment advice. Past performance does not guarantee future results. Trading involves risk of loss.