The same strategy. Opposite outcomes.

One set of rules — Tom Hougaard's opening-bar breakout with pyramiding — tested on two years of live broker data. On the Nasdaq it made +$15,400. On the DAX, the identical rule lost $9,300. Not a different setup. Not a different market regime. One number changed: the spread.

US30 · Dow · spread 0.7

+$15,400

ASRS, 3 tranches
PF 1.372 · drawdown $1,866

same ruleonly the cost differs

DE40 · DAX · spread 4.0

−$9,300

ASRS, 3 tranches
PF 0.61–0.72

Why this happens. Pyramiding pays the spread on every tranche. Three and a half tranches × 4 points = 14 points of cost per trade on the DAX, versus 2.5 points on the Dow. The strategy's raw edge is about 14% — smaller than the toll. The edge was never the problem; the toll was.

Watch an edge die

Our own 17-year School Run backtest on the DAX — 6,780 trades. Drag the cost and watch the profit factor fall.

Cost per trade

—

Profit factor

—

Net result

—

Edge remaining

—

Round-turn cost: spread + slippage, in points

—

Our own headline needed a correction

We published this School Run backtest as profit factor 1.35. That number is gross — it charges no spread at all. It is the honest measure of the signal, but it is not what you would have earned. Charged the DAX's real 4-point spread on all 6,780 trades, the same result is PF 1.145: still positive, but roughly 40% of the edge is paid away in toll. We've corrected the label on the backtest page.

The number that matters most isn't the profit factor at all — it's the break-even cost: the toll at which the edge hits zero. For this strategy on this data it is — points per trade. Above that, no amount of skill at picking entries saves you.

Same family of rules, tested per instrument

From a separate two-year study on live broker data (~550,000 five-minute bars, spreads measured on a real account, in-sample and out-of-sample split). These are per-instrument tests — unlike the slider above, which applies different costs to one DAX series to isolate the mechanism.

Instrument	Spread	Profit factor	Verdict

The FTSE is the exception that proves the rule. There, cost isn't the whole story: the strategy's gross profit factor was only 1.011 — no edge even before you pay anything. Our independent test of Tom's FTSE first-bar strategy reached the same conclusion. Some markets are cheap and workable; some are expensive; the FTSE was simply empty.

What to do with this

Check the toll before you trust the backtest. Take any strategy's average win and loss, count how many trades it makes, and multiply by your real round-turn spread. If that number is a large fraction of the gross profit, the strategy belongs to your broker, not to you.

Prefer cheap instruments for high-frequency rules. The more often a strategy trades — and the more it scales into positions — the more the spread decides the outcome. A rule that adds three tranches pays the toll three times.

Measure your own spread; don't assume it. The costs used here were measured on a live account during the actual trading session, not taken from a broker's marketing page. Spreads widen exactly when these opening-bar strategies trade.

Slider data: our School Run engine, DAX cash 5-min, Feb 2009 – Jun 2026, 6,780 trades, stop scaled to price, exit at the close (tools/spread_tax.py). Per-instrument table and the $15,400 / −$9,300 comparison: two-year study on live IC Trading broker data, Aug 2026, anti-lookahead protocol with in-sample/out-of-sample split. Strategy rules attributed to Tom Hougaard's public playbook; all analysis is ours. Educational content — not investment advice. Past performance does not guarantee future results.