I did a similar backtest on the Russel 3000 going back 15 years. The backtest suffered from survivorship bias which helps the OP's fallacy...
Rules were to go long when a stock was off 50% from its 252 day high, sell STOP =50% with a 100% profit target..
50% Stops were hit 57% of the time and the 100% profit target was reached 43% of the time.And that doesnt account for survivorship bias which probably would have skewed the results closer to 62% and 38%..
So while there is an apx chance of 38% that a stock doubles after getting cut in half,there is a 63% greater likelyhood that a stock off 50% declines another 50%...
In the spirit of full disclosure,the system was profitable,apx 5.9% per year,but did suffer a 60% max drawdown...
At the end of the day its a traders FALLACY to believe 50% stocks EASILY go up 100 percent.
Going down another 50% is the safer bet
Rules were to go long when a stock was off 50% from its 252 day high, sell STOP =50% with a 100% profit target..
50% Stops were hit 57% of the time and the 100% profit target was reached 43% of the time.And that doesnt account for survivorship bias which probably would have skewed the results closer to 62% and 38%..
So while there is an apx chance of 38% that a stock doubles after getting cut in half,there is a 63% greater likelyhood that a stock off 50% declines another 50%...
In the spirit of full disclosure,the system was profitable,apx 5.9% per year,but did suffer a 60% max drawdown...
At the end of the day its a traders FALLACY to believe 50% stocks EASILY go up 100 percent.
Going down another 50% is the safer bet
the math is the dubious part. it's a fallacy. I mean it is being interpreted as an indication of probability which is a fallacy.
To test this concept, I used data from this post (26 long ETFs for up to 23 years of daily data adjusted for dividends and splits with swing changes base on the highs and lows of at least the previous 8 trading days).
I found 6440 upswings with
mean time 13.3113354037267 trading days
median time 10 trading days
mean dollar change per share per trading day change 0.481707399701033
median dollar change per share per trading day change 0.238060333333333
And 6430 downswings with
mean time 9.53965785381027 trading days
median time 7 trading days
mean dollar change per share per trading day change 0.561237425952638
median dollar change per share per trading day change 0.27939325
The downswings happened faster with more movement per bar than the upswings.
So, this evidence demonstrates it's easier for prices to go down than up.
Last edited:

