Should Gap Moves Be Followed or Faded? (18:18)

By Lawrence G. McMillan

This article was originally published in The Option Strategist Newsletter Volume 18, No. 18 on September 24, 2009.

This is a question that has always intrigued me. We all see situations, for example, where stocks make a large gap move on earnings. Is it justified? Does the stock continue on in the same direction a week, month, or quarter later? Or does the knee-jerk reaction to news just provide a place for a reversal? These are all good questions, and so we have been working on a study to answer them.

It turns out that the study has raised even more questions. Does it matter if the gap move was due to earnings or to, say, an FDA event? How do broad market moves affect the results? Does the size of the gap move matter (that is, are larger-sized gaps more likely to follow through than smaller-sized gap moves)?

In this article, we’ll attempt to answer at least some of these questions, although this subject is so broad that another article may be necessary – at some time in the future – to present the complete data.

Definition of The Gap Moves Study

For the computer to be able to analyze this data, it was necessary to quantify what is a gap move. The following definitions were used:

a) from one day’s close to the next, the stock must have changed in price by at least 4 points or at least 10% b) in addition, there must have been a true gap1 on the chart of at least one point for stocks selling over $20, or 0.75 points for stocks selling between $10 and $20, or 0.50 points for stock selling under $10.

In other words, not only must there have been a sustained one-day move (supposedly generated by some news that surprised the fundamentalists and caused them to rush into or out of the stock), but that move must have also created a true gap on the chart. This means that we are excluding from our study any stocks that moved sharply on intraday news. But since we are attempting to isolate stocks that actually moved on news, this is the only way that we could do it without some massive database linking news to time of day moves in stocks. As it was, we had to manually link news on a particular day to stock moves on that day.

Once the gap moves were isolated, we still had to refine the data even more. For example, consider Sinopec (SNP), the China Petroleum and Chemical Corporation. This stock gaps almost every day, as shown in Figure 1. Obviously, this stock needs to be deleted from the study, for we are trying to isolate gap moves that are eventdriven and not just the normal course of affairs.

This was not an extremely difficult condition to impose on the data. If one particular stock gaps more than 4 or 5 times per year, it is probably a “chronic” gapper and should be eliminated from the results.

Finally, the remaining (qualifying) gap moves were sorted by the size of their first day move, in percentage terms. Then the results were analyzed in groupings of gap size. So, for example, a high-priced stock such as Google (GOOG) might gap more than 4 points several times per year. However, the percentage move would be rather small, so that GOOG move would fall into a “bin” of small percentage moves.

The initial day’s move – the one where the stock gaps – is, of course, not predictable. It’s what happens after that that is of interest. In other words, we assume that a trader would have had no way of knowing that the gap was going to occur and thus we are not assuming that one participated in the gap move. But if the stock gapped sharply to the upside, should one then buy the stock, figuring that the upward movement would continue on into the future?

We looked at three future dates: 10 trading days after the gap (two weeks time), a month after the gap, and three months after the gap. The stock price moves from the closing price of the “gap day” to the closing price of these future dates were compiled, and these percentage moves are what we are basing our analysis on.

Summary Of The Results

At first, we just analyzed the gap moves from 2009, figuring that we didn’t want too large of a data set to work with. It turns out that gap moves from 2009 – whether upside or downside gaps – all produced sizeable upward movements 90 days later. That seemed strange, of course, and obviously is a result of the fact that most of 2009 has been a roaring bull market. This demonstrates how the broad market can affect results.

So we analyzed the gap moves from 2008 as well. As you might expect, since 2008 was a bear market year, holding on for 90 days after a gap move almost always produced a downward move in the stock.

The one area, though, that was somewhat consistent in both years was that a large downside gap move generally does not follow through. Thus, a systematic approach of buying downside gaps does produce good results overall.

Results Detail

However, there is a lot of other information that can be gleaned from the data. For example, if we break out earnings-based moves from biotech news-related moves, does that alter the results? There are other reasons for gap moves, too: takeover bids, takeover rumors, lawsuit settlements and filings, and so forth.

But let’s start with the general results. The table below summarizes a lot of the data, based on the size of the one day percentage move created by the original gap move in the stock.

An upside gap is defined as the “open space” on the stock chart between one day’s high and the following day’s low. A downside gap is the “open space” between one day’s low and the following day’s high.

There is a lot of data in the above table, so let me first explain what the columns mean, and then we can delve into the result themselves. The first nine rows show results from 2009; the last eight rows are from 2008.

“% gap” identifies the size of the first day gap move, from close to close, as a percentage move. The top row in the table shows all gap moves higher of 50% or more in 2009. The second row, shows gap moves higher of 40% – 49%. The “–10%” row (purple cell) shows gaps moves lower of –10% to –19%. The “–30%” row shows gap moves lower of –30% or larger.

“# gaps” shows how many gaps fell into that category in that year. The next three columns (headings in yellow) are the average moves by the stock over the next 10 days, one month, and 90 days. For example, the grey cell, “–2.7%,” shows that the average stock that gapped higher by 50% or more in 2009 was actually 2.7% lower 10 days later.

The last six columns are actually two groups of three columns each. The three columns whose headings are in gold are a further breakdown of the stock moves after 10 days. “10 day % buys” is the percentage of times that the stock moved higher after the initial gap. For example, the red cell shows that in 2009, after 50% gaps higher, the stock moved higher over the next 10 days 41% of the time (you can infer, then, that 59% of the time, the stocks moved lower after the initial gap). Reading to the right, the average move upward was 22%, and the average move downward was 20% (the downward moves are shown as negative numbers, but in fact, had you sold the initial gap, you would have made 20%). In this case (50% gap moves higher in 2009), if you had bought every gap, 41% of the time you would have made money – 22% on average would have been your profit – but you would have lost money 59% of the time (losing 20% on average each time). The combined result of all those trades would be an average 2.7% loss (grey cell) as shown in the “Avg 10 days” results column.

The last three columns are a similar further breakdown of the moves over 90 days. In 2009, 67% of the time the stock moved higher after 90 days, and by an average of 71%! And, by inference, the stock moved lower after the gap 33% of the time, by an average of 35%. Those combined results are the average 35.7% move shown in the “Avg 90 days” column.

I have highlighted the “Avg 90 days” column in two shades of green. The lighter green shows the moves after 90 days in 2009. In every case, stocks were higher 90 days later. But is this because of the gap, or merely because the stock market was rallying strongly this year? You might point out that the market was down for the first three months of 2009, but gaps that occurred this year, when looked at 90 days later, were looking at a date after the March lows.

The darker green area shows the average moves 90 days after the gaps in 2008. Note that these are all negative. Again, was the bear market of 2008 more responsible for these moves than anything to do with gaps? When both years are taken together, it certainly seems like that the broad market movements had a dominant effect after 90 days. Similar results can be seen in the “Avg 1 month” column, too. In 2009, stocks were all higher one month later, but in 2008 they were all lower one month later.

Somewhat surprisingly, even the “Avg 10 days” columns shows the same effects – plus results in 2009, and negative results in 2008. The only row in the table that deviates from this pattern at all is the last one – gap moves lower by 30% or more in 2008. After those moves, stocks tended to rebound or at least remain flat.

So, in a broad sense, we can buy any gap (up or down) and hold for 90 days and have a profit, on average. At least in 2009. Now, I don’t think any rational investor wants to tie up the capital or the time to do that – after all, nearly 250 trades would be involved.

But in a general sense, this seems to point out that large downside gaps are often reversed to the upside after 90 days, even in bear markets. Each stock is different, of course, but one certainly should employ tight stops on any short positions held or established after the stock gaps lower by a large amount.

I recall that, in 2008, we were fortunate enough to own puts on some stocks that gapped sharply lower after earnings warnings were announced (those are sometimes tipped off by heavy put volume in advance of the actual corporate news event – which is the purpose behind our Daily Volume Alerts newsletter). We used tight stops and/or rolled those down to lower strikes, without necessarily having the benefit of the above study. But observations had taught us that sharp moves lower can quickly be reversed. In this instance, the data of the study presented here confirms one’s practical experience – something that doesn’t always happen with studies of this nature.

More Detail On Individual Stocks

It may also be important to understand why the stocks are gapping. In order to see this, we’d have to identify why each of the hundreds of gaps in Table 1 occurred. We don’t have that kind of time or manpower, but we did analyze the largest gap moves in 2009. The results are in Tables 2 and 3, which are printed on the insert to this newsletter (if you are reading a pdf document, they are on pages 13 and 14).

Consider Table 2 – the gaps higher. The six highest moves were a result of an FDA decision or the announcement of drug trial results by the company. This is why we follow the “Special Biotech Situations” in each issue of this newsletter, although the approach there is to attempt to buy straddles in advance of the move, hoping to take advantage of the gap move when it occurs. In this study, however, we are only considering the ramifications of where the stock goes after the gap occurs.

Moves due to earnings are noted in green. You can see that there are quite a few of those as well. In most cases, a move due to earnings does get some followthrough. However, there is always an exception to the rule, and one was State Street (STT) (bold line in Table 3). Even though the stock gapped down 59.1% on its January earnings report in 2009, it soon recovered and was 62% higher after 10 days and 147% higher after 90 days. We extracted the data for the big earnings moves from 2009. Without going into as much detail, the following table shows these results:

Earnings Big Gap Moves in 2009

.....Average result.....

So, if you want a short-term 10-day trade, you can buy up gaps on earnings and sell down gaps on earnings, on average, for a 10% profit. But after one month, there really isn’t any advantage at all. After 90 days, stocks were headed solidly higher – even those that had gapped lower – but some of that is certainly due to the strong market in 2009. But as was shown earlier, even in the bear market of 2008, large downside gaps were reversed to the upside, on average, after 90 days.

I did not have time to analyze the big moves from 2008. Also, my news database doesn’t go back that far for all stocks, so it is difficult to discern the actual reasons for the gaps. However, I will get that data and then construct tables similar to Tables 2, 3, and 4, for 2008 earnings moves. If that data confirms the results of Table 4, we may have a strategy that can be used.

Overall results – those shown in Table 1 – could be improved on a case-by-case basis. Consider the items marked in red in Table 2 (insert). They all have to do with takeovers, takeover rumors, or just plain unknown rumors. These are dangerous situations and the gap move is often reversed later on. Therefore, if one were to merely had avoided playing these in 2009, his overall results (assuming he was playing all the gaps) would have improved. You can see that – even in the the strong bull market of 2009 – all of these takeover-related upside gaps had turned back to the downside, after 90 days.

This article was originally published in The Option Strategist Newsletter Volume 18, No. 18 on September 24, 2009.