They took a sample of 25 stocks from their own well used watchlist (liquid, tradable stocks) and looked for large moves that produced gaps during the last 3 years. They were only interested in price moves that were larger than 1 Standard Deviation expected move (either up or down) which produced gaps in the charts of these stocks.
There's a lot of ways to calculate one Standard Deviation (I've written extensively about it already) but for this study, they looked at the closest expiration option chain, and took the ATM Straddle price (on the day before the gap) times * 85% = estimate of 1 Standard Deviation.
On this sample of 25 stocks, they discovered a total of 36 gaps that met the criteria of a greater than 1 SD move. Then the job was to see if the stock filled the gap over the next 30 days, and then over a 60 day period. (30 and 60 were used as those are the TastyTrade typical timeframe for trading options.)
I find the results astonishing. While traders walk around thinking stocks ALWAYS fill the gap (and of course they can and do sometime over a period of many months or years) but for a 30 to 60 day time period, they only fill gaps approximately 15% of the time. If you think your chances of that gap filling is 50-50, think again! Here's a recap of the actual data figures for a one standard deviation gap:
Okay, you might say, but what if it gaps down MORE than one standard deviation. Maybe a larger gap means a more certain retracement!? Tom's researchers thought of that too, so they continued testing. So they looked at 1.5 standard deviation, and 2.0 standard deviation move.
Compare the number of times the gaps filled, when the gap was even larger and you can see that the gaps filled even LESS often.
The 2 Standard Deviation move (a HUGE gap) produced the following statistics for this study, even smaller probabilities:
Correction: Both of the last two slides have typos on them, should show 30 days and then 60 days. (not 30 days/30 days)
This study shows 2 things:
1. The chances of filling that gap, in even up to 60 days, is less than 1 in 5.
2. As the gap increases in size, the likeness of retracement decreases.
Obviously this study does not mean that gaps might not fill on a different set of stocks in a different set of numbers. Obviously gaps DO fill, and you can easily look at a chart to prove it. But what's the probability? So this is not a NEW belief system to replace the old one that gaps always fill. What it is is a caution for you to stop and consider the probabilities before mechanically trusting such belief systems in your trading.
Sosnoff doesn't believe in trading by technicals or fundamentals. He believes in a probability/volatility game and to learn more of his amazing studies, see the archives on Tasty Trade.com.