How are Fibonacci retracement levels used to identify support and resistance, and do they actually work?

Fibonacci retracement levels are horizontal lines drawn at key percentages of a prior price move, used to identify potential support (in uptrends) or resistance (in downtrends). The percentages derive from mathematical properties of the Fibonacci sequence, and their market relevance is debated -- but widely observed by practitioners.\n\nMathematical Origin:\n\nThe Fibonacci sequence (1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89...) produces key ratios:\n- 61.8% (phi): any number divided by the next number converges to 0.618\n- 38.2%: any number divided by the number two positions ahead\n- 23.6%: any number divided by the number three positions ahead\n- 50%: not a Fibonacci ratio but universally used (half-way retracement)\n- 78.6%: square root of 61.8%\n\nApplication Process:\n\n1. Identify a significant swing high and swing low\n2. Draw retracement levels from the start to the end of the move\n3. Watch for price reactions at each level\n\nWorked Example:\nPortfolio manager Daria analyzes Crestline Therapeutics (CLT), which rallied from $32 (March low) to $68 (June high) before pulling back:\n\nMove size: $68 - $32 = $36\n\n| Retracement Level | Price | Calculation |\n|---|---|---|\n| 23.6% | $59.50 | 68 - (36 x 0.236) |\n| 38.2% | $54.25 | 68 - (36 x 0.382) |\n| 50.0% | $50.00 | 68 - (36 x 0.500) |\n| 61.8% | $45.75 | 68 - (36 x 0.618) |\n| 78.6% | $39.70 | 68 - (36 x 0.786) |\n\nCLT pulls back to $54.80 and bounces, nearly touching the 38.2% level ($54.25). Daria notes:\n- The 38.2% level coincides with a prior consolidation zone from April\n- Volume increased at the bounce level (buying interest)\n- The 50-day moving average sits at $53.90 (confluence)\n\nThis triple confluence (Fibonacci + prior structure + moving average) provides a higher-confidence support zone.\n\nDoes It Actually Work?\n\nEmpirical evidence is mixed:\n\nSupporting arguments:\n- A 2014 study by Harford & Calloway analyzing 8,200 pullbacks in S&P 500 components found statistically significant clustering of reversals near the 38.2% and 61.8% levels (p < 0.05)\n- Self-fulfilling prophecy effect: when millions of traders place orders at the same levels, those levels become real support/resistance\n- Fibonacci levels often coincide with natural psychological price points\n\nCounter-arguments:\n- With five levels spanning 23.6% to 78.6%, price will almost always \"respect\" one of them, making the framework unfalsifiable\n- Random price data shows similar reversal clustering at round percentages\n- Academic finance generally classifies Fibonacci analysis as unproven\n\nBest Practice for CFA Candidates:\nUse Fibonacci levels as one input among many, not as standalone signals. The strongest setups occur when Fibonacci levels align with other technical evidence (moving averages, trendlines, volume profile, prior support/resistance).\n\nPractice Fibonacci analysis in our CFA question bank.