The S&P/TSX Composite fell 629.00 points, or 1.74%, on Friday to close at 35,620.03, its worst session since June according to Bloomberg. The three sessions before it had added 476.91 points, taking the index from 35,788.18 on October 23 to 36,265.09 on October 28. One afternoon removed 152.09 points more than the entire rally had built.
The arithmetic is simple. The psychology is not. Friday was also the day the Nasdaq 100 dropped 3.5%, its deepest slide since early June, and Intel fell 29% after forecasting quarterly revenue below estimates, Bloomberg and Reuters reported.
Loss Aversion Weighs Friday at More Than Twice Its Size
Daniel Kahneman and Amos Tversky showed in their 1979 paper on prospect theory that people weigh a loss more heavily than a gain of the same size. In a 1992 follow-up, Tversky and Kahneman estimated the median loss aversion coefficient at 2.25. A loss feels about two and a quarter times as large as the equivalent gain.
Apply that weighting as an illustration, not a measurement. The Friday decline of 1.74% carries a felt value of about 3.9. The index had gained 2.9% in October through Thursday, so the felt loss exceeds the felt gain even though the month still closed 1.09% higher than September 30.
The TSX gained on 13 of the 23 sessions since September 29 and the average up day added 0.63%, yet the one session that closed 1.74% lower was 2.8 times that average.
Bars show the percentage change in each daily close from the prior session, and the shaded block marks the week of October 26, when the index closed 0.47% lower after three straight gains. Even after Friday, TMX Money data show the index up 12.321% year to date.
Daily Checking Turns Noise Into Losses
Shlomo Benartzi and Richard Thaler argued in 1995 that loss aversion becomes costly when it is paired with frequent evaluation, a pattern they called myopic loss aversion. An investor who looks at a balance every day sees a decline in 10 of these 23 sessions. An investor who looks once a year sees a TSX that is up 12.321% for 2026.
The behaviour that follows is documented. Terrance Odean analysed 10,000 brokerage accounts for his 1998 study in the Journal of Finance and found that households realized 14.8% of their paper gains but only 9.8% of their paper losses. That is the disposition effect: selling winners too early and holding losers too long. The winners sold went on to beat the losers kept by 3.4 percentage points over the following year.
Friday is the setting where that pattern is most likely to appear. Base metals and technology stocks led the TSX lower, according to MT Newswires, and the Nasdaq 100 completed its third straight weekly loss, according to Bloomberg. Recent winners were the holdings under the most pressure.
Monday Rebound Sets Up Recency Bias
Wall Street rose on Monday as oil prices tumbled more than 7% on hopes of a U.S.-Iran peace deal, Reuters reported, with the U.S. midterm elections due on Tuesday. Intel fell another 4.8%, the day after the worst one-day drop in its stock since the 1980s. Two vivid and opposite data points now sit three sessions apart.
Recency bias is the tendency to give the latest observation more weight than a longer record supports. After a sharp decline it tells investors that the selloff is the new trend. After a sharp rebound it tells them the danger has passed. Both readings come from the last headline rather than the full record.
The cost of acting on either reading is documented. Brad Barber and Terrance Odean, in a 2000 Journal of Finance study of 66,465 households from 1991 to 1996, found that the most active traders earned 11.4% annually against 17.9% for the market. With a binary event on Tuesday and a peace process that can reverse in a day, the next decision is likely to be made on whichever headline lands last.