Price Pause Strategy

[carltonp 10]

Hello Traders,

 

I'm trying to formulate a strategy that might take advantage of price pause after an instrument has paused.

 

Let me provide an example.

 

I have written a program that will alert me to when a price has paused for 5, 10 or 15 seconds. Also, the program will show me how many times a price has paused at the aforementioned seconds during, say a 5 min interval.

 

The strategy that I deploy using the formula is as follows:

 

YM (mini-dow) has reached an overbought/oversold stage on 5min chart. During the overbought/oversold stage the price pauses for 5 seconds at, say 11000. The price moves to 11005, and then 11008, but returns to 11000 and pause again for 5 seconds. The price then moves to various other prices but then returns to 11000 and then pauses for 10 seconds. This is all happening during the 5 min bar interval.

 

So, my strategy is to wait for a pullback on the following 5 min interval a enter at 11000 in the direction of the current trend.

 

So, my question is, can someone please tell me how I might improve on that stratey?

 

Alternatively, can someone please tell me if they deploy a strategy using price pauses in their trade setups? And any ideas would be great.

 

Cheers

 

Carlton

[wrbtrader 10]

Hi Carlton,

 

Most data providers have data pauses (stalls) due to bandwidth issues many times per trading day. These pauses or bottleneck of the data can last a few seconds. In addition, there's the issue of such occurring several times per day due to issues with the ISP.

 

Thus, with the combination of these data pauses occurring via data providers, ISP or computer issues...how will you be able to determine if it's a real price pause or an data stall via your data provider, ISP or computer connection.

[carltonp 10]

Hi Carlton,

 

Thus, with the combination of these data pauses occurring via data providers, ISP or computer issues...how will you be able to determine if it's a real price pause or an data stall via your data provider, ISP or computer connection.

 

HI wrbtrader,

 

That is a good question, and one that I probably could never answer with all certainty. However, a very polite engineer at IB kindly did some tests with me and for 30 mins we both synchronized on clocks and watched the pauses independently.

 

Having said that, I will never know for sure.

 

The question still remains: is this a strategy, worth pursuing in your opinion, if valid?

 

Cheers

[Tams 20]

you can do a latency test to see how many hops and the time delays between you can your broker.

 

there are free website that can do it for you,

also, IB's website has instructions.

[carltonp 10]

you can do a latency test to see how many hops and the time delays between you can your broker.

 

there are free website that can do it for you,

also, IB's website has instructions.

 

Thats cool. Tams, thanks.

 

I notice that nobody wants to stick there neck out and comment on the strategy - same with ET. Don't understand why....

[Tams 20]

Thats cool. Tams, thanks.

 

I notice that nobody wants to stick there neck out and comment on the strategy - same with ET. Don't understand why....

 

maybe that tells you something about the theory?

I don't know, I am open for ideas. Maybe the proposal needs more research and refinement?

Edited October 12, 2011 by Tams

[Tams 20]

you can do a latency test to see how many hops and the time delays between you can your broker.

 

there are free website that can do it for you,

also, IB's website has instructions.

 

how to use tracert

 

http://www.interactivebrokers.com/en/general/education/faqs/technical.php#trace_route

[carltonp 10]

maybe that tells you something about the theory?

I don't know, I am open for ideas. Maybe the proposal needs more research and refinement?

 

Thats interesting because traders are usually quick to comment on theories that are seen as not having any edge what-so-ever ...

[carltonp 10]

how to use tracert

 

http://www.interactivebrokers.com/en/general/education/faqs/technical.php#trace_route

 

Just applied the tracert command and the output shows a healthy connection....

[Tams 20]

Just applied the tracert command and the output shows a healthy connection....

 

You need to look at the number of hops.

the more hops between you and the broker,

the more chances for error.

 

when you are designing a strategy,

you have to take into account all the variables,

including those that are beyond your control.

[carltonp 10]

You need to look at the number of hops.

the more hops between you and the broker,

the more chances for error.

 

when you are designing a strategy,

you have to take into account all the variables,

including those that are beyond your control.

 

Hi Tams, just spent the last 45 mins watching todays recording. I have data feed coming in from both esignal and IB. I have checked every pause and compared them both and the pauses have all been in sync with each other.

 

Therefore, there is a problem it would be coming from my PC or my ISP, and I don't think there is any issues with either my PC or ISP.

 

Cheers

[Tams 20]

Hi Tams, just spent the last 45 mins watching todays recording. I have data feed coming in from both esignal and IB. I have checked every pause and compared them both and the pauses have all been in sync with each other.

 

Therefore, there is a problem it would be coming from my PC or my ISP, and I don't think there is any issues with either my PC or ISP.

 

Cheers

 

I am not saying there is a problem, or the will be a problem.

 

When you first try tracert, you reported everything is ok.

Of course things are ok. Why wouldn't it?

 

On your above test, you reported that the esignal and IB feed are the same.

Of course they are they same, why wouldn't they?

 

What you have observed is of very little value.

Let me explain.

 

 

Please take a look at the tracert report again. How many hops are there?

 

Mine has 14 hops.

Only the first 3 hops belong to my ISP.

ie. even though I have paid premium to subscribe to a fast internet,

the extent of my influence is only 3 hops !!!

the rest of the hops belong to other ISPs along the route.

If one of them has a hiccup, you are affected.

 

I am not saying someone will have a hiccup.

I am not saying internet is not reliable,

and I am definitely not saying your strategy is not workable because someone somewhere during the next one hundred years will have a 100 millisecond hiccup.

 

What I am saying is, you have to look at all the variables,

and see if you can control/depend on those variables.

 

Let's go back the the observation you made in the previous post.

Of course the esignal and IB data are the same...

if they go through the same hops, they will have the same latency -- ie. same results.

 

Next, you have to look at your current average latency,

and then divide that into your "Pause" Strategy threshold.

 

ie. you have to take samples of your tracert... to see how stable is your latency.

I have a 122 ms latency right now. It might be 60 ms tomorrow, or 250 ms in the next second.

You have to find out if you have a dependable varaible.

 

If the pause you are working on is 1 second,

and take my example, my latency is 122 ms (I will round it up to 200 ms, because you always round up),

then I would assume my expected error will be 20%. (200 ms of 1 second)

 

I don't know if it would work for you, maybe, maybe not. If it is acceptable, then by all means continue with the hypothesis. Otherwise even if it is a good theory, you are limited by our current technology.

 

Hope it makes sense.

Edited October 12, 2011 by Tams

[carltonp 10]

Hi Tams,

 

Thanks for the detailed response - really appreciate it mate. Below is the output from tracert - its not so great.

 

 

 

 

 

It totally makes sense about what you said about esignal and IB (don't know what I was thinking :-)

Edited October 12, 2011 by carltonp

[Tradewinds 12]

I notice that nobody wants to stick there neck out and comment on the strategy - same with ET. Don't understand why....

 

I'm looking at what I consider to be something similar to what you are describing. But I think I'm looking at a different side of the cube. (As opposed to the other side of the coin. Which implies something two dimensional.)

 

Personally, I think what you are doing is worth investigating. You have stated that you have found a way to know when there is a price pause. So you are capturing and analyzing intrabar price behavior. I'm doing something very similar, but I'm focusing on the price surge rather than the price pause, which, in a sense might be kind of the same as what you are doing. Maybe just the other side of the coin. Or a different side of the cube. :rofl:

 

Probably no one is commenting, because it's not a well known or widely used practice. I feel that those price pauses and price surges are a critical part to understanding price behavior. The objective is to find out how the market behaves and take advantage of it.

[carltonp 10]

 

Probably no one is commenting, because it's not a well known or widely used practice.

 

Thanks mate.

 

I will continue investigating....

[carltonp 10]

Hi Tams,

 

I wonder if you, or indeed any trader that trades YM, verify some prices for YM today?

 

As you're aware, I have been trying out a strategy that revolves around when prices pause.

 

Below is a list of times that the price of YM paused and the length of time it paused:

 

 

14-10-2011 15:00:58.58 Highlight 10 seconds, value =11536

14-10-2011 15:02:08.08 Highlight 10 seconds, value =11535

14-10-2011 15:03:25.25 Highlight 10 seconds, value =11533

14-10-2011 15:04:03.03 Highlight 10 seconds, value =11533

14-10-2011 15:04:39.39 Highlight 10 seconds, value =11534

14-10-2011 15:04:44.44 Highlight 15 seconds, value =11534

14-10-2011 15:05:19.19 Highlight 10 seconds, value =11534

14-10-2011 15:06:27.27 Highlight 10 seconds, value =11533

14-10-2011 15:06:44.44 Highlight 10 seconds, value =11533

14-10-2011 15:06:49.49 Highlight 15 seconds, value =11533

 

 

I'm sure there are traders that, like me, record their trades. Can you please look back and tell me if you noticed that the price paused at any of those times?

 

For example, the first trade on the list paused from 15:00:48 to 15:00:58 and so on....

 

This is really important, so you're help will be greatly appreciated.

 

Cheers

[Tams 20]

here's my data from IB (in ET time)

 

enjoy

ymz11 20110114-1500a.txt

[carltonp 10]

here's my data from IB (in ET time)

 

enjoy

 

Tams, I've been pouring over the figures line by line.

 

So, does the following mean the price paused at 11534 for six seconds? I guess not.

 

20111014,15:04:30,11534.000000,2

20111014,15:04:34,11534.000000,3

20111014,15:04:36,11534.000000,1

 

I'm trying to establish if anyone has observed the price pausing at the times I noted them to pause.

 

 

 

 

Cheers

Edited October 15, 2011 by carltonp

[uexkuell 10]

Some ideas, don't know if they could help:

 

- 5 min timeframe

In my view time based charts are not really helpful.

The timeframe is always chosen arbitrarily and bigger timeframes can lead to jumping into a move quite late.

Alternatives are volume based charts or just to look at pure ticks.

 

- Disconnect from quote feed provider / holes in data

I recorded many days of tick data from different data providers while constantly checking data integrity.

Sure there are sometimes holes in the data with every provider. But they occur quite infrequently.

Approaches like checking old (historical) data with other people probably will not lead too far because they cannot really establish confidence in future data integrity ("Past Performance Doesn’t Predict Future Results").

 

A suggestion for monitoring the data stream:

Subscribe to different symbols (some that show high liquidity / activity) that are not too strongly related. If the "price pause" occurs for multiple symbols at the same time there is a high chance of connection problems. Therefore the signal should be discarded.

If the pause is only for your target symbol not for the others it might be a valid pause.

For example if you want to watch YM, might choose CL and ZN to give this additional information.

[Tams 20]

Tams, I've been pouring over the figures line by line.

 

So, does the following mean the price paused at 11534 for six seconds? I guess not.

 

20111014,15:04:30,11534.000000,2

20111014,15:04:34,11534.000000,3

20111014,15:04:36,11534.000000,1

 

I'm trying to establish if anyone has observed the price pausing at the times I noted them to pause.

 

Cheers

 

There were no price updates,

don't know if the price paused.

[carltonp 10]

There were no price updates,

don't know if the price paused.

 

Hi Tams,

 

I'm afraid I'm not making myself clear. Looking at the sample of prices and times again that you provided below.

 

Between 15:04:30 and 15:04:36 the price was 11534 - that is six seconds.

 

My question is, did the price pause at 11534 For six seconds?

 

20111014,15:04:30,11534.000000,2

20111014,15:04:34,11534.000000,3

20111014,15:04:36,11534.000000,1

[Tams 20]

Hi Tams,

 

I'm afraid I'm not making myself clear. Looking at the sample of prices and times again that you provided below.

 

Between 15:04:30 and 15:04:36 the price was 11534 - that is six seconds.

 

My question is, did the price pause at 11534 For six seconds?

 

20111014,15:04:30,11534.000000,2

20111014,15:04:34,11534.000000,3

20111014,15:04:36,11534.000000,1

 

I think there is a misunderstanding of terminology here.

 

When I first read about your "pause" theory, I thought you meant there were no trading, thus the pause.

 

but you meant the trading was going on, but the price did not change. ie they were trading at the same price.

 

so yes, the price did not change between 15:04:30 and 15:04:36. There were 6 trades during that time, they were all consummated at 11534. Maybe they were traded at bid price, ,maybe they were traded at ask price, but they were traded at the same price.

[carltonp 10]

I think there is a misunderstanding of terminology here.

 

When I first read about your "pause" theory, I thought you meant there were no trading, thus the pause.

 

but you meant the trading was going on, but the price did not change. ie they were trading at the same price.

 

so yes, the price did not change between 15:04:30 and 15:04:36. There were 6 trades during that time, they were all consummated at 11534. Maybe they were traded at bid price, ,maybe they were traded at ask price, but they were traded at the same price.

Whoopee.

 

My program works. I reviewed every single price pause you provided and my formula was able to call it.

 

Cheers mate

[carltonp 10]

I think there is a misunderstanding of terminology here.

 

When I first read about your "pause" theory, I thought you meant there were no trading, thus the pause.

 

but you meant the trading was going on, but the price did not change. ie they were trading at the same price.

 

so yes, the price did not change between 15:04:30 and 15:04:36. There were 6 trades during that time, they were all consummated at 11534. Maybe they were traded at bid price, ,maybe they were traded at ask price, but they were traded at the same price.

 

It gives me the confidence to build on my strategy.

 

Thanks Tams

[Tams 20]

just for fun... here's the whole day of YMZ11

 

 

ps. no guarantee to accuracy or completeness.

ymz11 20110114-TICK.zip