Many scalping ideas look profitable on a chart and break even or lose after costs. The common reasons are structural, not bad luck.
Why strategies fail after costs:
Average winner is small compared with spread plus commission.
High trade frequency multiplies a small cost into a large one.
Stops slip in fast moves, so losers are larger than planned.
Entries at market during momentum pay the widest spread.
Characteristics of strategies that do survive:
Average winner at least four to five times the total cost per trade.
Fewer trades, focused on the best hours.
Limit entries where the setup allows.
Exits that avoid paying spread twice in quick succession.
A useful exercise: recompute your last 100 trades with costs doubled. If the result turns negative, the strategy is very sensitive to execution and needs a better cost structure or larger targets.
Has adding realistic costs changed your view of a strategy?
I find it clearer to express costs as a fraction of the stop rather than of the winner. With a 1.0 pip spread and 7 dollars round turn commission per lot, or 0.7 pips on EURUSD, every trade costs about 1.7 pips before slippage.
On a 6 pip stop that is 0.28R gone on every trade, win or lose. A system winning 55 percent at 1R makes 0.10R per trade before costs and loses about 0.18R after them.
The doubled cost test is a good stress, though I apply it unevenly. Limit entries get no extra slippage, while stops and market exits get an extra 0.5 pip, because that is where my real fills differ from the backtest.
Scaling out is a cost people tend to miss. Some accounts carry a minimum charge per fill, so on small lots closing half at 1R and the rest later can cost more than it earns.
Moving from 6 to 10 pip stops with targets scaled to match brought my cost down to about 0.17R. Fewer setups qualified, which was the point.
Has anyone measured how much of their edge disappears in the first ten minutes of London? My guess is the spread alone eats most of it.
One point in the list I would push back on a little is limit entries. They save the spread on paper, but the fills are not a random sample of the setups, because the orders that fill are often the ones where price kept going against you.
When I compared a limit 0.8 pips better than market on my own EURUSD trades, about 7 in 10 filled. The 3 in 10 that missed were mostly the clean runners, so part of the saving went straight back out as lost winners.
It also helps to run the numbers as a break even win rate rather than pips. With costs at 0.28R on a 1R target, you need roughly 64 percent winners just to stand still, against 50 percent with no costs at all.
Cutting costs by 0.1R moves that break even rate down about 5 points, the same as a filter that adds 5 points of hit rate. The cost side is usually the easier one to change, so that is where my testing time goes first.
Does anyone track the fill rate on their limit entries alongside how the missed ones would have done? I suspect most backtests quietly assume every touch fills.
Exotic pairs often attract attention because their ranges look large, yet the structure of their costs works against very short holding periods. Scalping depends on a favorable relationship between expected movement and transaction cost, and exotics frequently weaken that relationship.
Spreads are the first issue. Wider quotes mean a position begins with a larger deficit, and the move required just to break even is bigger. During quiet or off-peak hours, spreads in many exotics widen further, and quoting can become sporadic. Liquidity providers pull back quickly when conditions turn uncertain.
Swap and rollover charges are the second issue. Interest rate differentials involving exotic currencies can be large, so even modest holding periods across rollover may carry a noticeable cost or credit. This detail is easy to overlook when attention is on price alone.
Execution is the third. Thin books produce slippage, requotes and gaps, so stop levels may not be honored at the requested price. Price can also move in clusters, with stretches of inactivity followed by sudden jumps.
None of this makes exotics inherently unsound, but the cost, liquidity and gap profile suits slower strategies with wider targets better than rapid, tight-margin trading. Measuring spread against realistic range at the intended hour remains the clearest test.