BlueShip Research
Cost Is the Whole Result
Takeaway. A paper from CFM (Capital Fund Management) argues that trend at short horizons died because the market makers on the other side changed business models, and that tick size normalised by volatility decides which markets kept the effect. I transferred the test to daily US equities in a week. The predicted contrast replicated in direction at a t of 3.21 and died at a stressed 1.64.
1. Why a shop that runs on daily bars is reading a microstructure paper
In July 2026, Kurth, Eisler, Rej and Bouchaud of CFM (Capital Fund Management, a Paris-based quantitative hedge fund) presented "Is Trend Still Your Friend? A Microstructural Account of the Demise of Short-Term Trend-Following." Fast trend following went flat around 2009. The paper rejects crowding and proposes a change of counterparty. Market makers who warehoused inventory were replaced by HFT firms on intraday mandates to hold inventory flat, which will not absorb persistent directional CTA flow. The discriminant is tick size normalised by volatility. Trend survived in markets with large ticks and died in markets with small ones.
My engine tests signals at macro horizons on daily bars, so what kills my results may live in a resolution I cannot see.
2. The transfer, and the caveat I declared before running it
US large caps all tick at a penny, so tick size does not vary. The ratio of tick to dollar volatility does.
Ranking the panel monthly by that ratio produces two tiers, median annualised volatility 29.3% in the tier with the high ratio and 35.5% in the low.
I wrote the fatal caveat into the ticket before running anything: the paper's mechanism is intraday futures microstructure and this is a daily-bar equity transfer, so a clean failure is expected. Four looks were declared, continuation and reversal in each tier, setting the family bar at 2.96.
3. What happened
All four declared looks failed. The best reached a t of 0.10.
| Declared look | NW t | Stressed t | Bar |
|---|---|---|---|
| Continuation, high ratio | −3.40 | −4.86 | 2.96 |
| Continuation, low ratio | −5.72 | −5.95 | 2.96 |
| Reversal, high ratio | −1.33 | −6.95 | 2.96 |
| Reversal, low ratio | 0.10 | −6.39 | 2.96 |
The four looks are not the paper's prediction, which is a contrast: continuation relatively stronger in the tier with the dense grid, charged nothing extra because both looks were already declared.
The contrast produced +6.14% a year, Sharpe 0.66, Newey-West t of 3.21, clearing the deflated bar of 2.96 in the direction the paper predicts.
The rerun for margin of safety took it apart. Halve the gross, double the costs, and the t falls to 1.64 against the base bar of 2.50. Sharpe degradation from in-sample to out-of-sample is 80%, 0.89 to 0.18. Turnover runs about 100x a year, so cost is the whole result.
Rejected. A result that only exists at zero cost is not a strategy.
4. What I take from a rejection
The direction replicated, days after publication, on a variable that had never been in my engine. The tradeable version did not survive, and the honest reading is not that the paper is wrong. Its mechanism lives at a horizon where my costs are fatal and my data is blind.
5. The boundary, stated precisely
What I can test on daily bars: cross-sectional signals rebalanced daily or slower; conditioning variables constructed from price, volume, filings, and options open interest; realistic cost sensitivity through a stressed rerun; whether an effect concentrates in a handful of names or a single regime.
What I cannot test, and do not pretend to: order book dynamics, queue position, adverse selection at the tick, the shape of impact within a day, or anything that requires knowing what happened between the close and the open. My cost model is an assumption in basis points on turnover with a doubling stress, a discipline rather than an execution model. A real slippage model needs tick data and a venue map, and I have neither.
What this rejection tells me about that boundary: the one result derived from microstructure that I could get to the deflated bar died on turnover, the term a cost assumption built on daily bars estimates worst. Signals whose edge lives closest to the trading day are the ones I should trust least.
6. What would change my mind
The cheapest falsification: if only turnover kills the contrast, slowing the tier rebalance from monthly to quarterly should preserve most of the signal and cut most of the cost. That test is not run. I am naming it rather than doing it, and this note is worth less than it would be if I had.
Method. Panel of current US large caps, daily adjusted closes. Ratio defined monthly as tick over the product of price and return volatility over the trailing 63 days, ranked across the cross section and split at the median. Books are decile long-short within tier, weighted equally, lagged by one day, rebalanced monthly through the engine's standard backtester. Statistics by Newey-West with the family bar deflated for four declared looks; the stressed series halves gross and doubles costs per the house construction. The universe is survivorship biased, so every raw figure is an upper bound. Source paper: Kurth, Eisler, Rej and Bouchaud, CFM, July 2026.
Educational research only. Not investment advice.