Looking for and testing short term patterns in the market

[overit]

having questions like this help me to qualify why I do what I do, which helps me to spend a little time thinking about things more, which is never a bad thing

Thats what it is about. You come up with a thesis and test it on the market. It will let you know pretty quickly how wrong you are. Most of what I have read on the net has been quickly disproved by my testing. Doesnt matter who thinks your a dill if the market tells you, you are the forex king! LOL!

What you essentially have to do is impose a set % move up or down as a cutoff for the test in order for the test to accurately have a 1:1 ratio. This opens up an extra set of variables to any test done that can greatly influence the result.

On most of my pattern testing this is why I use a moving average on the ATR. As volatility goes up so does your ATR. Its not perfect but it tries to adapt to changing conditions. But in saying that when I use individual bar size as R/R it hasnt changed the results.

quite a few of the patterns I've tested in the thread so far have shown differing levels of performance over different timeframes

Yes and I believe there is good reason. I now trade 1hr+ timeframes even against my wishes. Prefer the action of the shorter stuff but have come to the realization that the bigger stuff trends better. Also I think because the stops are wider out and less trades you are better protected against noise. The ratio of spread to your profits is also lessened when taking bigger trades. ie 2pips is a lot when you take 10pips profit. Not so much when you take 100pips trade.

Something else I believe is that winrate is inversely effected by risk/reward. ie if you are chasing 1:3 you will drop back to 33% and you chase 1:4 you go back to 25%. (based on random entries). So really the 1:1 probably not a big deal. This gets to the crux of the matter, is it your pattern that is causing price to head in a certain direction. Or is it just being "fooled by randomness". (I recommend reading the book of that title too).

 

[Wysiwyg]

This gets to the crux of the matter, is it your pattern that is causing price to head in a certain direction. Or is it just being "fooled by randomness".

Very good.
I have yet to find a pattern that consistently results in a price rise or fall and I aint chasing my tail as much these days with Ami. back testing. Though I don't see the thousand plus back tests I have performed as a waste of time. They have shown me that patterns aren't what makes profitable trading.

 

[tech/a]

Hi guys
Interesting discussion.
Ive had sometime to wade through the chat.

Just a couple of my observations with regard to patterns (I use for Index Trading a couple of VSA patterns)

I have yet to find a pattern that consistently results in a price rise or fall and I aint chasing my tail as much these days with Ami. back testing. Though I don't see the thousand plus back tests I have performed as a waste of time. They have shown me that patterns aren't what makes profitable trading.

I strongly believe its the way you apply a pattern trade/s to your trading more than the pattern itself that will determine greatest success.

I only take pattern signals when I know the direction of the trend in which the pattern occurs.
If I get a reversal pattern signal long or short I find waiting for confirmation of a continuence in the alteration of trend is a must.
How do I determine trend.
Simply a line if its broken and there was a signal before it then I look for an entry at the break so I can set a trade on a Stop limit order.
I can stop and reverse with this method.
So I dont take signals against the trend without a break in trendline.
Works well.

The next is timeframe---how long is a pattern valid---my answer to that is until another pattern succeeds or Fails.
Here is a current FTSE Index chart marked up for the last few days and yes I'm still in the latest trade shown.



The next consideration is liquidity.
I have found that The FTSE which trades around 20X more contracts than the SPI is far more reliable and has far more available trades than the SPI.
Infact I cant see myself trading the SPI unless I see something that looks like Everest there to trade!
A 20 min high volume bar in the SPI is around 500 contracts
in the FTSE or DAX its 6000

Enough to ponder hope it is of some help.

 

[tech/a]

The above chart is the DAX
Brain fadded when labelling I have both open--trying to do too much at once.

 

[overit]

I strongly believe its the way you apply a pattern trade/s to your trading more than the pattern itself that will determine greatest success.

Totally agree.

I have a pattern that will make you laugh at its simplicity. Does it read the markets mind or punch above it weight?... I really doubt it although it is the best pattern I have tested. What makes it so great in my eyes is that it is frequently occurring and happens regularly at short term tops and bottoms. It gives me an entry point and set risk. The hard part is working out when to exit.



Both of these patterns are identical in my eyes.

I dont know if this pattern has a name. My theory behind the pattern is once you have 3 higher lows you put a sell order under the third low (as per pic) and keep following the lows upwards until you have a break down. And inversely for lower highs you put a buy order on the third lower high. I particularly like this pattern when forming double dip bottoms or tops and on the "wrong" side of the moving average (against the trend *ducks for cover!*) and also on short term S/R.

 

[overit]

Here is a few basic examples of the above pattern in action from the current euro 1hr action. The pattern is nothing special but notice it forms regularly on reversals getting you into the trade. Also something I havent mentioned is the times that these patterns are more likely to be reversals. Around london open, us open and london close are great times and to a lesser extent asian open. Makes sense when you think about it. Daily high and lows can be great for your support and resistance zones if in doubt. Obviously the bigger S/R zones the better chance of a solid bounce but sometimes you just dont know. After all that said and done the pattern is just and entry. You need to find your exit mojo as that is where your money is made or lost. If you go for the big outliers prepare for more losses and bigger drawdowns. This is where the peter crowns and threalthing links from earlier come into play.



Here is something I did to check out the main reversal times using this pattern. (Times in MT4 time). The bulk of the action is in the london session.

 

[overit]

This chart is a good example of what I mean by exit. Most of these trades went to 1:1 but there was one massive outlier and another decent size one. If you waited for the big trade you would have carried a fair bit of a drawdown (if discarding the use of a break even stop.) If we took 1:1 we would have been going fine but would have missed the eventual smoking hot run loading our bank and buying that gold ferrari! LOL!

 

[Wysiwyg]

This chart is a good example of what I mean by exit. Most of these trades went to 1:1 but there was one massive outlier and another decent size one.

Matey those black background .jpg images make it hard to see the chart drawings.

Also your RRR of 1:1 is particular to you. This ratio will be different for most of us as we have different risk.

 

[overit]

That last chart was more of an example for the exit principle of cutting it short or letting it run rather than my entry points. If you want a clearer view it is the last 2 weeks from the euro.

Here is an interesting post from hanover at the forex factory. It is great food for thought. It also makes you question your trading strategy and where you edge lays. Some good points.

With this pattern I have presented I feel that the "edge" is not the pattern but trading the rebound off support and resistance and the "outlier" runs that it produces. The pattern is just something to outline my trade setup. Commonly when S/R breaks it just blows straight thru and not producing a pattern as such.

The pattern could be easily described as a break in lower highs (downtrend) when coming to support or a break in higher lows (uptrend) when coming to resistance.

With this theory you could use pinbars, engulfing candle inside bars or a myriad of other cool indicators which would all be telling you the exact same thing.

I really like inside bars also. I trade the setup on a 4hr chart. It tells you a similiar thing. 4hours of ranging and a break of the high or low. But you cant get too carried away. I have MT4 charts and IG charts. They differ by 1hour in their time zones. Guess what you get different bars for same action. What I learnt from this was that the IB usually occurs in the asian hours and the break out is the european opening. Kinda fits in with what I said before about major reversal times. Again its just an entry signal and defined risk. Only half the work is done.

 

[Wysiwyg]

Totally agree.

I have a pattern that will make you laugh at its simplicity. Does it read the markets mind or punch above it weight?... I really doubt it although it is the best pattern I have tested. What makes it so great in my eyes is that it is frequently occurring and happens regularly at short term tops and bottoms. It gives me an entry point and set risk. The hard part is working out when to exit.

Overit, I would like to see the results of a couple of these patterns in the now if possible. If they are frequently occurring then do you have any to observe the outcome over the coming week?

 

[overit]

Overit, I would like to see the results of a couple of these patterns in the now if possible. If they are frequently occurring then do you have any to observe the outcome over the coming week?

Draw a line off fridays high and low and we will be perched waiting for the week to start. I dont like to carry trades through the weekend.

Now the reason I think this is my best pattern is not because it punches above its weight. Its because of the frequency of where it happens. (S/R, head fakes, double bottoms/tops, etc). Thru my testing period of 2009 the raw pattern (no filter) punched around 53% for 1:1. Stop was ATR of entry bar or ATR(50) whichever was smaller. Good but the first pattern I posted was the same with a much larger statistical base and it bummed out. I stopped testing this pattern when I realised a greater calling for it. (I have to do this manually so I only test what looks the best).

This is an enlarged chart from the last 2 days on the eur/usd 1hour. As you can see the raw pattern happens a lot. They are very easy to pick yourself. Notice what I said about the times also. Asian session and when the US is closed is ****ty. Money, news moves the market.

(click on image to enlarge)


If looking to use this for your own entries I would recommend to take it around levels you think it will rebound off. Support/Resistance, overbought/oversold, trendlines etc. Personally I only use horizontal S/R that are only a few days old and havent been broken. I have a few other little nuances but thats your prerogative. This is how I tested it and the results were great. It works now I trade it live but not to say it wont stop working sometime in the future. The nature of the market! I am always looking for new ways to trade the market.

To complete this trade is your call. I use a 50sma type setup. Peter Crowns in his DIBS thread has an interesting method of using 1hour entries (he uses inside bars) to build structures that ride the daily trend. What your hoping is that when the price goes back to test the support or resistance it rejects it, you nab the entry and it continues the trend or starts a new one.

 

[Wysiwyg]

Draw a line off fridays high and low and we will be perched waiting for the week to start. I dont like to carry trades through the weekend.

So it is a short sell at 1.2908? There is old resistance at 1.2912.

 

[overit]

So it is a short sell at 1.2908? There is old resistance at 1.2912.

I think this one would be a personal judgment. Yes it is the raw pattern. I also think new higher resistant point weakens the older ones, it happened out of hours and I wouldn't trade it coming into a weekend. Could gap either way. I would just wait for a pattern to form in our blue areas in the new week. I have a theory on the weekend gap and Asia's ability to push new ground but it needs further testing.

The last trade was the green circle. It traveled 130pips before turning around. It was a trade because it occurred level with the previous day high (resistance). You could also make a case for the bottom support zone on that last flick up but I wouldn't have as it was a bit far away from my zone. Personal preference.

My last actual trade I placed was the setup on the left of screen. Triple top at the highs and I landed a good trade down and had the rest of the week off.

 

[zipzap]

Awesome posts professor!! I trade futures in much the same way!

 

[professor_frink]

Awesome posts professor!! I trade futures in much the same way!

Thanks zipzap, appreciate it!

A big welcome to the forum too. Hope you enjoy your stay

 

[professor_frink]

afternoon folks

Haven't had too much to add this week, seems to be a case of steady she goes!

Just having a quick look at what happened last night. First thing that stands out is the daily range has been quite low of late(to be expected coming out of a corrective period though).

To have a look at things a little closer, I've tested out how the market has generally behaved in the past based on similar circumstances.

Test below is looking for price coming off a 20 day high(which we made yesterday), today's range is in the bottom 10% of readings for the past 3 weeks:



Not really much to get excited about either way, but isn't really suggesting much weakness coming through yest, even after the recent rally.

Just to get a slightly more specific picture, I filtered the results based on today being a lower close than yesterday and on lower volume than the day before(which is what we've seen happen overnight):



This does weaken the results a little, but the results don't really suggest anything that would be of a concern to anyone holding longs right now

Having a look at other instances of this type of behaviour on the SPY, it's a bit of a mixed bag - happens fairly randomly and at various stages of trend, probably not too much that can be read into it either way. Looks like waiting for the market to develop a little from here is the way to go.

Cheers

 

[Wysiwyg]

afternoon folks

To have a look at things a little closer, I've tested out how the market has generally behaved in the past based on similar circumstances.

Test below is looking for price coming off a 20 day high(which we made yesterday), today's range is in the bottom 10% of readings for the past 3 weeks:

Hi Prof, Knowing this is testing short term patterns I would like to suggest that your sample period doesn't reflect the statistics as much as a maximum sample period would. There are many nuances in price movement and they are terribly hard to quantify or exact. Especially over short periods. What do you think about that?

 

[professor_frink]

Hi Prof, Knowing this is testing short term patterns I would like to suggest that your sample period doesn't reflect the statistics as much as a maximum sample period would. There are many nuances in price movement and they are terribly hard to quantify or exact. Especially over short periods. What do you think about that?

I think I'm not quite sure what you are getting at Wys! Could you clarify it for me a little more?

 

[Wysiwyg]

Here is a few basic examples of the above pattern in action from the current euro 1hr action. The pattern is nothing special but notice it forms regularly on reversals getting you into the trade. Also something I havent mentioned is the times that these patterns are more likely to be reversals.

Unfortunately this week on this particular occasion the pattern did not manifest and the longer term up trend continued.

I know where you are coming from though.

 

[Wysiwyg]

I think I'm not quite sure what you are getting at Wys! Could you clarify it for me a little more?

How far back is your back testing sample period Professor?