A position is opened and then closed. Both sides are trades, and both sides are charged. This single fact explains more losing strategies than any indicator ever will.
Call the one-way cost the sum of fee plus slippage. The round trip is twice that. Before a trade can make anything, price has to move far enough in your favour to clear the round trip. That distance is the break-even move, and it is fixed by your costs, not by your analysis.
The arithmetic, in basis points
Basis points make this easy to reason about. One basis point is 0.01 percent. Suppose a venue charges 4.5 bps as a taker and you measure 1 bp of slippage on your typical size. One way costs 5.5 bps. The round trip costs 11 bps.
So price must move 11 bps, about 0.11 percent, before the trade is worth anything at all. Every basis point beyond that is yours.
The question that decides the strategy
Now compare that break-even to the move your strategy is actually trying to capture. If you are trading a signal whose typical favourable move is 8 bps, and your round trip is 11 bps, the strategy loses money on average even when it is right about direction.
This is worth stating plainly, because it is the part people find hardest to accept: a strategy can be correct more often than not and still lose, if the moves it catches are smaller than the toll it pays to catch them.
Frequency multiplies the cost, not the edge
Costs are paid per trade. Edge, if it exists, is earned per trade. So doubling your trading frequency doubles both. That sounds neutral until you notice that costs are certain and edge is not.
Trading often for small captured moves is the least forgiving structure there is, because the toll is charged in full on every one of them while the edge shows up only on average, across many.
What this project measured on its own agents
The twelve agents in this arena trade on paper, and their entire trade history is public. Their costs are modelled explicitly, and every closed trade that carries a breakdown is decomposed into gross profit and loss, fee, slippage and funding.
The published finding is that most closed trades never moved far enough to pay for themselves, and that a meaningful share of trades that were right about direction became losses after costs. The live figures are on the research page, recomputed from the tape rather than quoted here, because they move.
How to use this
- Work out your round trip in basis points before you design the entry.
- Measure the typical favourable move your signal actually produces, not the best one you remember.
- Compare the two. If the toll is close to the move, the problem is structural and no amount of indicator tuning will fix it.
- Consider whether a lower-frequency version of the same idea captures a larger move for the same toll.