Crude Oil Trading Results and "Dead Time"

[karl9000 10]

I've been trading CL using SST since it came out in August. I've been trying various iterations of a "trade plan" using UTA as my analysis tool. I'm not the most confident of traders no matter what system I used. But in looking at both live trade results and backtesting (I'm a LOT more "confident" in backtesting! ) I've noticed something and I really need to know if it's just me or if there is "something" there.

 

Here it is: whether it's live trading or backtesting, I get poor/crappy results from the 0912 to 0930 time interval (exchange time). Whether I lock in one tick at MM or use the Channels as my stop, I just don't produce positive results. In one testing scenario, I actually have a winning percent number for that time period, but negative Net$!

 

Admittedly, this may be due to my temerity in trading and adjusting my entries; I'm not as "nerveless" as TJ and others! But even in my recent backtesting where I don't move my stop to BE at MM as aggressively as before, I still end up with that time period as a net money loser.

 

Does anyone else see this in their data? What's hurting me is that even though I'm no longer trading that time window, I'm still counting winners during this period towards POQ. It's been realt tough this week since if I didn't count them torwards POQ, I could have been more profitable than I am. But I want to make sure it's not just me and that others are seeing it before I commit to not counting these towards POQ.

 

Thanks!

[foxstamp 10]

I've probably done more backtesting of CL with the SST than anyone (with the only possible exception being TJ himself). You are correct in seeing that the time period immediately before the stock market opening does not have results as profitable as other time periods. I believe it is because a setup may occur because of what crude oil is doing, however once the stock market opens crude oil follow stocks for a while and that can cause the CL setup to follow stocks instead of its own setup. Just my thought. ??

 

I don't have my trading computer with my UTA with me today and I need to cut this short as I need to help family members move to a new house this morning. But I saw your message and wanted to give at least some input.

 

I am in the training room and I take all CL 377 trades (except I don't go against an extremely oversold/overbought market as determined by the Netpicks MACD). I don't want to step in front of a moving train!

 

My POQ is 2 wins and Positive or I stop with 3 wins, whichever comes first.

 

If you'd like to ask me anything my email is: jon@foxstamp.com I'd be happy to help in any way.

 

Good Luck

 

Jon Zeininger

[jdevron 10]

Time based charts can have dead zones. Tick based charts are designed to alleviate this problem. A 1597 tick chart would complete one bar every 1597 ticks. A 60 minute bar chart would complete one bar every 60 minutes.

[karl9000 10]

I didn't menton in my original post but I am trading the 377 tick CL chart.

[tjnoon 10]

Thanks for starting this thread Karl. It is very satisfying to see that there are traders out there using the UTA in the way it was intended AND deriving valuable information out of their trade data as a result. I have included a screen shot of the TimeofDay Summary studies that are a part of UTA so that those reading this thread can actually see what we are talking about. The yellow fields are where you can type in a particular time and the following fields autmatically adjust to create the 'time window.' I put a few more screenshots up just so that everyone can see the same thing I am seeing in relationship to this thread.

 

Since the SST came out, following the tradeplan(s) as posted on the Owner's Club webpage (there are two variations for the 377 tick), I too am seeing a negative result during the time period you are referring to. Whereas 8:50 to 9:12 we have enjoyed 68.5% winners, the next 18 minutes the win rate dropped considerably, as did the net result.

 

While Foxstamp has an interesting theory, I can't say for sure why this is happening. I will say though that there really is not enough data during that time peroid to come up with any longer term conclusions. The jury is still out even if it is not looking good at the moment.

 

There were a total of 59 trades during that time period. Remember, we are entering two positions for each trade and the UTA sees each position as an individual trade. You have to keep that in mind when you are looking at the numbers. So while it appears there were 118 trades, you really need to divide that number in half. From what I have learned during my use of the UTA is that you really need anywhere from 75 to 150 trades to establish somewhat reliable stability in your data. So for that reason, I would say the jury is still out.

 

I went back and looked at my 610 tick trade data from last year, end of Sept to end of Jan 2009. This was the actual trade data that I used to develop the SST. During that time, the 9:12 to 9:30 time window produced 96.4% winners! BUT, there were only 27 trades during that window. Also, it was a 610 tick and not the 377. Karl, perhaps you might want to look at the 610 but it is ironic that the losingest time period lately for the 377 was the winningest time period a year earlier, for the 610 tick.

 

A couple more comments: First, based on the current 377 evidence, it seems to me it would pay to avoid trades during that time period. I would go back and backtest it. Let the results of your backtest tell you whether you should use any winning trades during that time window towards your quitting goals (poq, 'power of quitting') or not. I suspect that if you have a plan that does not include that window, not applying them to your quitting goals feels right to me. Just a hunch though and I could be wrong. What is the edge of your tradeplan? If it doesn't include that window of time, then why include it in your quitting goals unless you can prove you need to.

 

Second, take a look at the attached equity curve. Notice how flat the equity was from about mid Oct. to mid Nov. We really could not get any traction during that time range yet the equity still managed to hang out near to its high water mark levels before finally breaking out to the upside in a big way. Maybe go back and look at the trade data to how many trades during the 9:12 to 9:30 'dead zone' affected that sideways action in the equity curve. Remember, with UTA you can easily look at different sub sections of your data using the Enter Rows feature, A23 and A25.

 

For those who don't have UTA, I can tell you this. During that sideways equity curve, the winniest window of time, 8:50 to 9:12, produced a 58.3% win rate while the losingest window, 9:12 to 9:30, produced 100% winners! Huh? Lol.. I don't mean to befuddle you with this information but it says to me that markets have the tendency to phase shift and sometimes a period of time will underproduce for a while. Then a phase shift will occur and the underproducing time will catch up and begin to over produce while another time will begin to underproduce. Understand the metaphor? The markets are infinitely challenging! Paying attention to phase shifting will help you interpret price action in a different kind of way.

 

As an old musician (drummer) I actually did learn something from my guitar player friends when they connected all their electronic gadgets to their electric guitars. I just didn't realize what I had learned until I started noticing this 'phase shifting' phenomenon within market price action. lol.. See if after another 40 to 60 trades +/-, the 9:12 to 9:30 stats remain what they are, or not. By then, there should be enough data to be able to cautiously hang your hat on.

[wout 10]

Hi guys,

 

Here we are over a year later and I did my back test on the UTA using Trend Jumper and lo and behold I too found this period of under performance on my dataset from August 2012 to December 2012 using a 377 tick chart.

 

Interesting comment about the phase shift and the 610 tick chart not having this issue.

 

Also the flat spot on the equity curve calls for a reversal system. The TJ is a continuation system and it may benefit from a system that can offset it in those times when that may be advantageous. I am working on one.

 

My live trades are doing well now and I have started to produce my own upsloping equity curve.

[tjnoon 10]

Ha, Funny! I haven't been on this forum in too long a time, which isn't meant as a commentary of what I feel about this forum because I think it's great, but I have been a bit preoccupied with other projects. Nice to see Wout coming in here and reviving what could and should be an important thread.

 

I'm sure many on this forum aren't too familir with Trend Jumper but just to respond to Wout's comment about needing a reversal setup, I would just suggest using the crossover trade and turning off the ema filter on that setup. Team it up with some sort of overbought/oversold indicator like the OSOB that comes with SST or some other histogram, like what Foxstamp suggested. His use of the MACD with the custom settings could work well, or a number of others. That would give you your reversal setups when the odds are at their best. Trend Jumper does have that ability but you have to move beyond the vanilla training which is meant to cater to a wide range of trader skill levels, as far as a training program goes. Also, you could use my idea where I just measure the distance between the EMAs in relation to the Jumpline and the price combined with good 'ole fashion chart pattern recognition; failed swinghighs/lows, the rolling over the EMAs a la Trend Jumper, etc..

 

Cheers!