Trading is a competitive, mostly zero-sum game played against professionals, minus fees. Real edges exist, but they are small, they decay as others find them, and costs eat a large share of gross returns. No strategy below “works” unconditionally — each one is a specific bet that wins in some market regimes and loses in others. Anyone selling a strategy with guaranteed returns is, without exception, selling something else.
The strategy families at a glance
| Strategy | The bet | Wins when | Bleeds when |
|---|---|---|---|
| Trend following | Moves persist | Sustained trends | Choppy, sideways markets |
| Mean reversion | Extremes snap back | Ranging markets | Strong trends (extremes extend) |
| Breakout | New highs/lows continue | Volatility expansions | False breakouts in ranges |
| Grid | Price oscillates in a band | Sideways chop | Sustained one-way trends |
| DCA (investing) | Asset rises over years | Long horizons, rising asset | The asset simply doesn't recover |
| Arbitrage / carry | Price gaps between venues close | Fragmented, hot markets | Competition compresses spreads; venue risk hits |
| Market making | Spread > adverse selection | Calm two-way flow | Toxic one-way flow (news, crashes) |
Trend following
The oldest documented systematic approach: buy what is already rising, sell what is already falling, cut losers fast, let winners run. Its plausible foundations are behavioral — information spreads gradually, investors herd, and winners attract flows. Its cost is psychological and statistical: most trend trades lose small amounts in chop, and profitability depends on a few large winners paying for many small losses. If you cannot tolerate being wrong 50–60% of the time, you cannot run trend following. A common refinement is trading pullbacks within a trend — waiting for a dip against an established trend before entering, to avoid buying local tops.
Mean reversion
The opposite bet: price stretched far from its average tends to snap back. It produces the seductive profile of many small wins — until a real trend arrives and the “stretched” price keeps going. Without hard stop-losses, one bad trade can return months of profits. Mean reversion without an exit plan is how accounts die slowly, then suddenly.
Grid trading — and its hidden skew
Grid bots (the standard offering inside exchange apps) place a ladder of buys below and sells above the current price, harvesting oscillations. In a sideways market this prints a stream of small profits. The catch is the payoff shape: negative skew. Many small wins, and then a sustained trend fills every buy level on the way down, leaving a large losing inventory. The frequent-small-win pattern makes grids feel safer than they are — the risk is concentrated in the rare event, exactly where intuition fails.
DCA: investing, not trading
Dollar-cost averaging — buying a fixed amount on a schedule — is not a trading strategy; it is a disciplined way to accumulate an asset you have independently decided to hold for years. It removes timing decisions and emotion. Its honest limit: it contains no risk management and no exit. If the asset never recovers, DCA just averaged you into a loss more smoothly.
Arbitrage and market making
Price gaps between exchanges, futures-vs-spot basis, and funding-rate carry are real, measurable edges — which is precisely why professional firms with colocated servers, fee rebates and 24/7 infrastructure compete them down to slivers. Retail attempts usually discover that after fees, withdrawal delays and the occasional frozen venue, the “free money” wasn't. These niches reward infrastructure, not insight.
Systematic vs discretionary
Any of the above can be traded by feel or by fixed rules. Rules have three honest advantages: they remove in-the-moment emotion (the biggest documented destroyer of retail returns), they can be tested against history, and they produce a record you can actually evaluate — trade by trade, cost by cost. What rules do not do is guarantee the edge itself: a backtest is a look at the past, easily overfit to it, and markets change regimes. Automation executes a strategy consistently; it cannot make a bad strategy good.
What realistic success looks like
A workable retail approach is unglamorous: a small edge, applied consistently across hundreds of trades, with strict risk control so that no single trade or losing streak is fatal, judged on data over months — not on last week's PnL. Most tools traders lean on for entries are indicators; what they can and cannot deliver is the subject of the next guide.