The most popular category of EA is the scalper: many trades a day, a few pips each, a win rate in the nineties, and a balance curve that rises every day. It is the dream of trading as a salary. It is also the category in which almost every product fails, and the reason is not the strategies. It is arithmetic.
A trade's cost is the spread, plus commission, plus slippage, paid every time, win or lose. On a major pair at a good broker that is around a pip. A scalper aiming for three pips pays a third of its potential in costs before it has done anything, and at an ordinary broker, or at the hours it trades, it pays more. A strategy that holds for a week and aims for eighty pips pays the same pip and barely notices it.
Frequency multiplies the problem. A scalper taking ten trades a day pays its costs two and a half thousand times a year. A weekly system pays them fifty times. For the scalper to end the year in the same place, its edge per trade must be fifty times larger relative to costs, and no retail strategy has that.
The high win rate is a symptom, not a strength. A scalper closes winners fast and stops out slowly, so most trades are small wins and the occasional loss is large; or it does not stop out at all, and the losses accumulate as open positions until the account ends. Either way the win rate is manufactured by the exit rule and says nothing about the edge.
The reality is slower and less exciting. Strategies that survive live trading tend to trade rarely, hold for days, aim for moves large enough to dwarf their costs, and lose four trades in ten. Their records have losing months and no rising line to admire each morning. They also still exist a year later.