Trading The Opening Gaps

[jasont 10]

In my journal I made this post in regards to research I did on the opening gaps. It was recommended that I post a thread in here to get the alternative views on the topic as it has probably been investigated by people here before. Here is the post below with my research:

 

As I have been noticing some big moves at the open I thought that it might be a good idea to see how I could capitalize on these moves. At first I thought there was a connection between where the Tick opened and where the market subsequently traded for the morning.

 

Following that basis I decided to research the Tick and the open to find what the real result was. I managed to get 33 days of data to form my analysis on. It began from 17th March 2008 and went to 6th June 2008. Its not a massive amount of data to use but I'm not forming a mechanical trading system. I simply want to find an edge that can be used in conjunction with my interpretation of the market.

 

During that period there was 21 negative Tick opens versus 12 positive ones. On 20 days the Market initially moved in the direction of the Tick and 13 times the market moved against the Tick straight out of the gate. That was promising until I looked at the values the market moved. When the market moved in the direction of the Tick it averaged 3.21 points. The average move the market made against the Tick before going the direction of the Tick was 1 point.

 

This was ok but I didn't like the ratio. I looked further into different relationships between whether the market opened positive or negative and if the Tick followed suit. After some searching I came to opening gaps.

 

I found that 22 of the 32 gap days we closed the opening gap. My definition of a gap was having an open 0.25 above or below the previous close. Closing of a gap I defined as reaching the high or low of the last 5 minute bar the day before. I thought this was interesting as it sort of went against the Tick relationship yet still had a 60% success rate.

 

I would have suspected that had we opened down we would get a negative Tick reading and vice versa. To close the gap would be going against the Tick reading.

 

Now I had a mentor that traded the close of an opening gap and mentioned some days were better than others and the size of the point gap also mattered. I have found that Wednesday and Thursday were the most reliable to close the gap. I also found that a gap over 8 points would rarely be closed. In fact only 1 in 6 gaps over 8 points were closed. If I removed these 6 big gaps from the data, we would only have 26 gap days and 21 of those were filled. Now that seems like a decent edge.

 

Concerned as some of these gaps were only small, I wanted to find out what the size of these moves were. Considering the 3-1 ratio the Tick idea presented I wanted something pretty solid. So I recorded the size of the move that closed the gap and also recorded the size of the initial pullback prior to closing the gap. So say we gapped down 4 points to 1390 and moved down to 1389.25 before moving up to 1396.50 and coming back. The initial pullback size was 0.75 and the size of the move that closed the gap was 6.50.

 

The results were pretty darn good. I found the average size of the move when we closed the gap was 5.60 points whilst the average size of the initial pullback was 0.67 points. The biggest move when closing the gap was 12 against the biggest pullback of 1.75. The lowest move when closing the gap was 2 points whilst the lowest pullback was zero.

 

Now if I limited this to only trading gaps below 8 points I would possibly make a profitable trade of 1.75 points on 21 trades whilst losing a maximum of 1.75 points on 5 trades. I decided to look further at which days tended to close the gap more than others. For specific weekdays, the ones which closed the gap most commonly were Thursdays and Fridays. The ones which performed the worst were Mondays and Tuesdays. Funny as these were the same days my mentor mentioned to me a long time ago.

 

Of the 6 Mondays, 3 closed the gap. Of the 8 Tuesdays, 3 closed the gap. Of the 6 Wednesdays, 6 closed the gap. Of the 8 Thursdays, 7 closed the gap. Of the 5 Fridays, 3 closed the gap.

 

From here, if I removed the Tuesdays and Mondays as well as gaps above 8 points, these are how my stats pull up.

 

Total Gap days excluding Mondays, Tuesdays and moves above 8 points: 17

 

Total Days Gap Was Closed: 15

Ratio of Days Gaps Closed: 88%

 

Average Size Of Gap Move: 5.95

Average Size Of Pullback before Gap Move: 0.70

 

This has been very interesting. There was no strong relationship between upward and downward gaps that were reliable. I will now look at paper trading this though from these numbers this seems very much a good way to trade some of the opening days. The ideal way to do this would be getting a trade as close to the open price as possible with a stop 1.5 points away. The first profit target would be placed at 1.75 level and then using the second position to trail.

 

Although this limits me to only trading the open during the second half of the week, it does keep the edge in my favor.

[MC 10]

Interesting...subbed.

 

The only part I don't get is the 8 point or larger...this is the ES I presume and not the YM correct?

[jasont 10]

Yeah sorry MC. The gap is for the ES. What does subbed mean?

[MC 10]

  jasont said:

Yeah sorry MC. The gap is for the ES. What does subbed mean?

 

Subbed means...Subscribed to see what you or others report on gaps.

 

So from your studies middle of the week is the time to fade it seems.

I'm curious of any psychology behind the gap moves if you guys have thoughts.

 

On stocks they use gaps to keep you out of the move often, not really the case here. I know gaps on indices tend to be like black holes when weakness is shown and they gravitate to filling the open gaps. Kind of like how virgin POC tend to be tested at some point.

 

Just throwing some random thoughts out. :missy:

[jasont 10]

Hey MC thanks for the replies and interest.

 

Now I am not sure why Wednesday and Thursday tend to be the better days. Maybe the market gets an idea of fair value later in the week. I will try to outline my thoughts on the psychology behind it as best as possible but think others will come up with some good ideas also.

 

I think the market is ruled by the big players. Lets say we look at yesterdays range. We began a decline in the afternoon. Some of the big players may not have been willing to get on until we saw a breach of the 1350 area. If we gap down to the 1350 they will be banking on one of two things happen. We go beneath the 1350 support area or we bounce off it.

 

If they see a breach of the support coming, they want to get the best position possible. The big players begin buying up light and selling bigger to the following thinking the market is bouncing. Trying to get the best possible entry. Market goes up and closes the gap. Then as we come back down and breach the support they can start unloading their big positions on the way down as average Joe trades the breakout move.

 

If they see a bounce from support coming they may buy straight from the gate or have worked up their position during the pre market. They may even try to push the market down a bit over the support area to buy the breakout traders. The ES commonly tests the support and resistance levels to run a few stops and gain position.

 

Just some ideas behind the psychology I came up with. Hope to see others with ideas as to why this may occur or may not continue to occur.

[firewalker 10]

Nice idea to open a thread for...

 

I thought I'd just copy my original reply from your blog here, just in case anyone else want to follow the discussion with these things i mind:

 

Several comments:

 

(a) [starting a new thread, which has been done]

 

(b) Don't forget to take into account the news that often is released 30 minutes after the open.

 

© Traditionally there are four types of gaps: common, breakaway, continuation and exhaustion gaps. See also here: http://stockcharts.com/school/doku.p...ool:glossary_g

 

(d) What might be more significant in trading gaps, is to determine whether or not the gap means a break of important S/R. This might help in determining the probabilities of a move in the same or opposite direction.

 

(e) In futures because of the overnight trading there is technically no gap as everything is pretty continuous and markets can be traded 20 hours out of 24...

 

I was wondering if anyone else played with gaps? Because in general, they do have a tendency to get filled...

[firewalker 10]

  MC said:

So from your studies middle of the week is the time to fade it seems.

I'm curious of any psychology behind the gap moves if you guys have thoughts.

 

 

Although I've never given it much thought nor attention, I have read (and observed some things in my -relatively short- trading career):

 

Fridays then tend to be difficult days, especially after lunchtime.

Fridays tend to be trend-following (it was especially true last year, when we had the parabolic rise). In bear markets, Fridays then to be the days when not a lot of people want to be long...

 

To determine whether there is any correlation with gaps being filled or not on certain days, I think you'll need a whole lot more stats than just 20 days for that.

[jasont 10]

FW those are some interesting findings on Fridays. I think what I find most interesting is that for a long time it has also been said people tend not to hold long positions over the weekend. I wonder if that is a phenomenon brought on by the news during the weekend prior to 1987. Maybe people have decided the weekend risk is too high in general?

 

I agree that 20 days is hardly enough to go by for solid analysis. I would love to have at least 50 days analysis for this type of thing. Unfortunately I don't have access the the backdated data. If anyone is willing to do a check, I guess mainly of Wednesday to Friday I wouldn't object.

 

I guess for me the most interesting part is of 32 gap days we had 22 gaps closed. That alone is a high ratio of 68%. There has been a long time belief that gaps had a tendency to be filled in the market but only now have I done any solid research into it. If we include the gaps from Monday and Tuesday this week, Wednesday didn't gap in my analysis, we have 24 of 34 gap days filled. Bumps the ratio up to 70%.

[firewalker 10]

  jasont said:

I guess for me the most interesting part is of 32 gap days we had 22 gaps closed. That alone is a high ratio of 68%. There has been a long time belief that gaps had a tendency to be filled in the market but only now have I done any solid research into it. If we include the gaps from Monday and Tuesday this week, Wednesday didn't gap in my analysis, we have 24 of 34 gap days filled. Bumps the ratio up to 70%.

 

Well I had a quick look around the web and found some stats for January 2008 on the DOW. They came up with very similar numbers actually:

 

"Of the 20 trading days in January, 13 (65%) of those days produced gaps in the DIA, and of those 9 (70%) of those gaps filled for a profit using the classic “fade the gap” strategy."

 

http://blog.afraidtotrade.com/gap-fade-stats-for-january/

[jasont 10]

That is a very interesting find FW. It seems this thread really hasn't caught the attention of many which is unfortunate because I would have liked to hear from the experience of others. Even as MC suggested, the psychology behind why gaps might have a tendency to close.

 

What I also found interesting about the research done on the site you linked to said Wednesday and Thursday saw no gaps made that weren't filled. Adding further weight to the theory that Wednesday and Thursday appear more likely to close. Something definitely worth investigating in my opinion.

[OAC 10]

Jsont:

May be it is a good idea to examine all the gaps that are greater than 8 pt., and see how many get 50% of the gap filled ? When the gap is large, 50% is worth shooting for.

[jasont 10]

Thanks OAC, that is a great idea. I'd imagine on a large gap there would be some type of pullback that some of the big traders would be pushing for to get a good size position should we continue in the direction the gap was made. If the odds do favor that then a scalp for the first position could work and then trail the rest for the possibility the large gap closes.

[LiggerPig 10]

Hi Jay,

For such a small sample size you've done well to find some important ideas you can trade successfully. It's very true that opening gaps of 7 points or less tend to be filled and 8 points or over tend to lead to runaway, "stuck up" or "stuck down" trend days where pullbacks are shallow and the moves are difficult to join in.

For someone, like yourself, who is actively watching the opening, it's very useful to know how the pit operates to maximise their profits.

They take advantage of those who have a bias at the open, those who must have a position in the market, whether it be due to news, data releases, etc.

It's an emotional time when, very often, one side wants to trade regardless of slippage. The pros on the floor want to suck in both sides and they want each side to pay over the odds. It's one of those times these guys make their big money, moving the market one way and then the other.

Importantly, you may be able to find a way to successfully trade the gap-fill with a decent win and profit ratio. You will probably need to set a time to enter e.g. I've used 9.45 am for longs which meet certain conditions such as 20ema on 15 min chart is pointed up along with a second, confirming signal on a smaller timeframe.

I'm sure you'll find some conditions / rules in your methods which give you an edge.

 

You'll need to paper-trade the ideas since, unfortunately, your current stops are probably far too tight to trade gap-fills. But trading the gap-fill remains a very very popular method of making money at the open. It's not my style but it works for others.

 

I'm more nervous at the open than at any other time, and it's not by accident that I will very often note the ES price at 9.30 ET. A tight opening range in the first couple of minutes becomes a 'pivot' which shows up later. You notice it more at s/r levels. I think Fiewalker touched upon something similar in your original thread?

 

Best to note how price moves away from your s/r levels, rather than trying to second guess a turn, i.e. if price falls to your support level, don't just enter long at support. Rather, recognise the opening ranges as a time when one or other side feel they "must" get in a position at any cost, whereas you can wait for a pullback. You'll achieve a safer, low-risk entry and easier, less volatile exit by picking them off on your terms.

 

You have to distance yourself from the opening emotional swings, One of the toughest things to accept about trading the open is you will get stopped out only to see the market go your way immediately afterwards. This will happen sometimes day after day and is why you must be selective and only enter at prices which offer a good risk/reward ratio. Scale out at on a partial gap-fill and don't hold out for the full gap-fill if it co-incides with your s/r levels.

 

Finally, many trading systems were developed and became very popular over the years based on breakouts from the opening range. You should already be very aware that your favoured market deliberately creates false breakouts in the form of marginal new highs and lows. Bear this in mind when trading gap-fills!

[jasont 10]

Liggerpig first off welcome aboard the TL forum and I must say your first post is superb!

 

Thank you very much for that great post, it has a wealth of information in there and some excellent advice. You are spot on about trying to keep a tight stop at the open and like FW said, finding something to support the closing of the gap.

 

The insight into the pits is a valuable addition to the psychology seen at the open. Like yourself I am probably most nervous at the open for some reason. I do constantly wonder why that is considering I never hold an overnight position. There is no pressure to place a trade, maybe it is somehow related to the emotions felt prior to a game of sport or competition of some sort.

 

I personally am better at trading after an hour or so of market action because I can piece together parts of the puzzle so I may not end up trading the gaps at all. I will however paper trade it to get a better understanding.

  Quote

Best to note how price moves away from your s/r levels, rather than trying to second guess a turn, i.e. if price falls to your support level, don't just enter long at support. Rather, recognize the opening ranges as a time when one or other side feel they "must" get in a position at any cost, whereas you can wait for a pullback. You'll achieve a safer, low-risk entry and easier, less volatile exit by picking them off on your terms.

 

You just seem to be coming out with the great advice on this post! I often find these are the safer entries as opposed to catching the market directly at the s/r levels. The only downside is when we have a runaway market move substantially first.

 

  Quote

Finally, many trading systems were developed and became very popular over the years based on breakouts from the opening range. You should already be very aware that your favoured market deliberately creates false breakouts in the form of marginal new highs and lows. Bear this in mind when trading gap-fills!

 

Hehe, this is exactly why I can't bring myself to trade a breakout system on the ES. Even when the market passes my s/r levels I still have caution about a false breakout with the need for a too greater stop level.

 

Thanks again Liggerpig. You have provided me with some great looks into the opening gaps. In one way the odds are stacked in the favor of the person wanting to fade them. I guess it just depends on the person whether they can put the faith in the system alone.

[firewalker 10]

  jasont said:

FW those are some interesting findings on Fridays. I think what I find most interesting is that for a long time it has also been said people tend not to hold long positions over the weekend. I wonder if that is a phenomenon brought on by the news during the weekend prior to 1987. Maybe people have decided the weekend risk is too high in general?

 

 

I did some research myself and was surprised to hear that there is more to the so called calendar effect than just anecdotal evidence. Apparently there is statistically significant evidence that shows that the market does have a tendency to react in a specific way on certain days or times of the year.

 

Not all of these 'calendar effects' hold up after a statistic test. The October effect for example, seems more psychologically - because of the crashes in the past - than anything that actually holds up over time.

 

What was interesting about all of this, is that the Friday Effect has been subject to a recent study by Leon Zolotoy (attached). In his conclusion he states:

 

Our results strongly suggest that the firms continued to report "bad" news on Fridays during the last two decades. The mean earnings surprise is significantly lower and the proportion of "bad" earnings announcements is significantly higher for the Friday earnings announcements compared to other trading days. This tendency persists over the whole time span of our study.

 

Moreover, our findings suggest that for the last ten years the investors systematically overreact to the "bad" earnings announcements released on Fridays, compared to their response to the "bad" news released during other trading days.

 

I've also attached another paper (Dellavigna & Pollet) which has some interesting findings:

Friday announcements have a 15% lower immediate response and a 70%

higher delayed response. A portfolio investing in di?erential Friday drift earns substantial abnormal returns. In addition, trading volume is 8% lower around Friday announcements.

[jasont 10]

That is indeed an interesting and useful find Firewalker. Thanks for the great reference material mate. So did you find this solidified your own findings for a Friday? I still wonder if the bad news reactions on Fridays has any correlation to the 1987 crash. Many long time traders I have spoken to said a few events over that weekend had them thinking "oh no" before the open on Monday.

[firewalker 10]

  jasont said:

That is indeed an interesting and useful find Firewalker. Thanks for the great reference material mate. So did you find this solidified your own findings for a Friday? I still wonder if the bad news reactions on Fridays has any correlation to the 1987 crash. Many long time traders I have spoken to said a few events over that weekend had them thinking "oh no" before the open on Monday.

 

1987 is over 20 years ago now, and I think since then a lot of new market participants have entered who don't think or reflect about these kind of things. Although the authors talk about the effects being more psychologically than statistically verifiable, I think it's important to be aware of some calendar effects which have shown up after time consistently.

 

Michael Carboni (http://www.livewithoscar.com) for instance never trades the day before a holiday weekend because he has reason to belief -and as he's been trading for 26 years, so this is obviously based on experience in the field- that market participants behave differently. I think I read somewhere that Fridays (especially after lunchtime) most of the pro's have left the field...

[Guest forsearch]

  firewalker said:

 

Michael Carboni (http://www.livewithoscar.com) for instance never trades the day before a holiday weekend because he has reason to belief -and as he's been trading for 26 years, so this is obviously based on experience in the field- that market participants behave differently. I think I read somewhere that Fridays (especially after lunchtime) most of the pro's have left the field...

 

This guy, known as "Oscar", is also charging pikers in Australia $50 a round-trip for his services. I couldn't believe it when I saw it on his Aussie seminar, so I've attached it below.

 

Just go to the 1hr 17min mark and you'll see it for yourself.

 

 

Fifty (yes, 50) dollars a round trip. No typo or misquote here.

 

Unbelievable.

 

-fs

[firewalker 10]

  forsearch said:

This guy, known as "Oscar", is also charging pikers in Australia $50 a round-trip for his services. I couldn't believe it when I saw it on his Aussie seminar, so I've attached it below.

 

Fifty (yes, 50) dollars a round trip. No typo or misquote here.

 

Unbelievable.

 

-fs

 

The only reason I referenced to him was because he mentioned trading (or not trading) on specific days... $50 is definitely a lot, but as I'm not sure about what his services include or not, I'll refrain from giving any comments. Besides, this is hardly the thread to do so... I only know of him because of the free videos he puts out, but some parts of the videos smell like pushing people into buying something. He says the reason why he stopped offering some things for free, is because too many people were acting on the signals he gave out...

Edited June 18, 2008 by firewalker

[Guest forsearch]

I've added my comments onto a new thread here:

 

Trading the OMNI live with O$car

 

Feel free to post your comments there.

 

-fs

[jasont 10]

UPDATE:

 

After two weeks of recording the market from my initial research I have an update. The figures aren't as great as they were but they are definitely are still high probability.

 

I now have a total 43 days I have looked at with 40 of those 43 days being gap days. 27 of those days have seen their gaps filled. That gives a 62% of gap days being filled. Thats not fantastic but a good trader could probably trade a tight stop on these and look for the runaways when the turn is sensed.

 

Looking further into my 8 points theory we have now got 30 days of data where the gap was 8 points or below. 24 of the days where we gapped below 8 points were filled. That makes 80% of gap days below 8 points filled. That is more like the figures I am looking for.

 

Now if we only include Wednesday, Thursday and Friday in the gap days that are below 8 points we have 19 days of data and 16 of those days were closed. That gives us a slightly better result of 84% success. It's better than the results including all days with a gap beneath 8 points but it may not be worth leaving those other days out.

 

I decided to look into the risk reward ratios associated with each type of gap day to see if there was any major differences.

 

If we included all gap days we have an average gain of 5.93 points when closing the gap and an initial average pullback of 0.78 points. That itself is quite good considering 62% of the gaps are closed.

 

If we only used gap days of 8 points or less we have an average gain of 5.47 points when closing the gap and an average initial pullback of 0.80 points. That looks much better when we include the 80% success rate.

 

If we narrowed it again to only the gap days 8 points or less on Wednesday, Thursday and Friday we have an average gain of 5.56 points with an average initial pullback of 0.80 points. That is pretty much the same as the above results.

 

That leads me to believe leaving out Monday and Tuesday is not worthwhile in trading the opening gaps. When trading gaps below 8 points on everyday you have 30 possible trading opportunities as opposed to 19 when limited to Wednesday, Thursday and Friday.

 

There are 30 gap days below 8 points of a total 40 gap days. This means a trade occurs on 75% of gaps. However there are only 19 gap days below 8 points on a Wed, Thurs and Fri of a total 40 gap days. This means only 47% of the gaps can be traded. The extra 4% success rate on these trades doesn't make up for the extra 30% trade opportunities, especially when the risk reward is very much the same.

 

I'm going to take a more serious look at this by paper trading the gaps and seeing what happens. It could be very interesting to see what occurs.

 

No Of Days Calculated: 43

No Of Gap Days: 40

 

No Of Days Gap Filled: 27 (Average gain 5.93 points, Average pullback 0.78)

Gaps below 8 points: 30

 

Gaps below 8 points Closed: 24 (80%) (Average gain 5.47 points, Average pullback 0.80)

Gaps below 8 points on Wed, Thurs, Fri: 19

 

Gaps below 8 points on Wed, Thurs, Fri Closed: 16 (84%) (Average gain 5.56 points, Average pullback 0.80)

Edited June 22, 2008 by jasont

[firewalker 10]

  jasont said:

There are 30 gap days below 8 points of a total 40 gap days. This means a trade occurs on 75% of gaps. However there are only 19 gap days below 8 points on a Wed, Thurs and Fri of a total 40 gap days. This means only 47% of the gaps can be traded. The extra 4% success rate on these trades doesn't make up for the extra 30% trade opportunities, especially when the risk reward is very much the same.

 

I'm going to take a more serious look at this by paper trading the gaps and seeing what happens. It could be very interesting to see what occurs.

 

Nice work jason. One remark though: 40 days is already more representative than 20, but perhaps it'd be interesting to see how this plays out in an uptrending instead of a downtrending market... You could go back to last year and just select a number of consecutive days and see how the stats play out there.

[jasont 10]

  Quote

Nice work jason. One remark though: 40 days is already more representative than 20, but perhaps it'd be interesting to see how this plays out in an uptrending instead of a downtrending market... You could go back to last year and just select a number of consecutive days and see how the stats play out there.

 

Thanks Firewalker, I agree that it will be interesting to see what occurs in regards to the gaps on an up trending market over a longer time frame. What is interesting though is that I have recorded how many days trended up in the morning, how many trended down and how many saw a ranging market.

 

The criteria was a general trend from 9.30am to 12pm. Of 43 total days we saw 20 days trend up for the morning, 13 trend down and 10 range for the morning. Not quite the same as seeing an up trending market but it is interesting that we had more days where the market trended up in the morning. I haven't assessed how many of those 20 up trending and down trending mornings saw the gap closed but it looks like I will need to get those stats together.

[reres 10]

very interesting stuff, i have been studying gaps recently alot. does anyone have some more info on tick readings and other indicators to trade the gaps

[scottzim46 10]

Does anyone here have an indicator that shows the cash and futures close with dotted lines and also calculates teh half gap fill autuomatically on .d charts every morning automatically after the cash opening?