Technical analysis is the study of price and volume patterns to predict future prices. Its reputation in academic finance is poor (efficient-markets theory says it shouldn't work). Its practical track record is mixed (some patterns work in some regimes; many don't work at all). The honest take: technical analysis is neither voodoo nor magic — it's a tool, useful in specific contexts and dangerous when applied uncritically.
Two foundational ideas — both real
- Trend-following: prices that have moved up tend to continue moving up (momentum). Statistically supported across asset classes and time periods. Foundation of CTA strategies.
- Mean reversion: prices that have moved far from a moving average tend to revert toward it. Works on short horizons (intraday) and very long horizons (decades) but weakly in between.
These two ideas are in tension. The market is sometimes trending (riding momentum is profitable) and sometimes ranging (fading extremes is profitable). The skill is identifying which regime you're in. Most retail technical analysis confuses the two and underperforms.
Common indicators and their honest assessment
- Moving averages: smooth price series. 50-day and 200-day moving averages are watched by many traders, so they have some self-fulfilling-prophecy support. Useful as trend filters.
- RSI (Relative Strength Index): measures recent gains vs recent losses on a 0-100 scale. Above 70 = overbought; below 30 = oversold. Used for mean-reversion signals. Works in ranging markets; fails in trends.
- MACD (Moving Average Convergence Divergence): a momentum indicator. Heavily watched; weak statistical edge. Useful as one input among many.
- Bollinger Bands: 20-day moving average plus/minus 2 standard deviations. Mean-reversion bands. Useful for short-horizon trading; less useful at longer horizons.
- Volume indicators: rising prices on rising volume confirm a trend; rising prices on falling volume suggest the trend is exhausting. Real signal here.
What doesn't work
Most chart patterns (head and shoulders, cup and handle, ascending triangles) have been statistically tested and show weak-to-no predictive power. Elliott Wave theory is essentially unfalsifiable. Fibonacci retracements have intuitive appeal but no statistical edge. Reading these in retrospect is easy; predicting them in advance is much harder than chart-pattern enthusiasts admit. Senior traders use them as one input among many, not as primary signals.
Support and resistance — the one universal
Price levels where previous activity clustered (a recent high, a round number, the close before a major news event) tend to act as support (prices bounce when reaching them from above) or resistance (prices stall when reaching from below). This works partly because many traders see the same levels and place orders there, creating self-fulfilling clusters. Even strictly fundamental traders use support/resistance for execution — 'don't sell below 17.40, that's a major support level.'
When technical analysis matters most
- Execution: even if your trade idea is fundamental, the price you enter matters. Buying at support gives better risk-reward than buying mid-range.
- Risk management: stop-loss placement often uses technical levels. A stop just below a major support is logical.
- Position sizing: trades with cleaner technical setups can warrant larger sizes than ambiguous ones.
- Trend identification: simple moving averages tell you whether the market is in an uptrend, downtrend, or range — useful regime context for any strategy.
Exercise
You're long a Kenyan equity at KES 100. The stock has rallied from KES 80. Major moving averages: 50-day at 95, 200-day at 85. Recent support cluster: 92-95. Walk through: where would you place a stop? Where would you take partial profits? What would change your view?