Price Action Strategies All Break Down

[Swan Noir]

I'm in total agreement with dbphoenix. What the hell do you think PA is? I just don't get guys that need to break it down to some simple equation that they believe is backtestable for a million years.

Context, including (in a major way) S/R, is KING. I have a really sweat setup that is currently treating me very well in CL. But I don't even look for the setup until I see either the bulls or bears strongly in control and I know I want in. The setup is, for me, only a mechanism that gets me into a trend at a sweat spot more often than not. I'm loooking for minimum MAE and a good entry bar. Does it mean the trend will continue? Certainly not but it means that when my read on context is right I have a nice positive expectation. That's what I seek.

 

[Sergio77]

Expected return from random-walk is negative if you include trading costs. Your time-frame is random-walk. ALL patterns in a random-walk are ...random.

 

[gmst]

Quote from Sergio77:

Expected return from random-walk is negative if you include trading costs. Your time-frame is random-walk. ALL patterns in a random-walk are ...random.

According to you, what time frame is not random walk??

 

[Dhalsim]

Quote from Sergio77:

Expected return from random-walk is negative if you include trading costs. Your time-frame is random-walk. ALL patterns in a random-walk are ...random.

Okay i don't want to argue whether markets are random or not at this timeframe as you most likely are correct.

However, can you explain to me then how something (which is not indicator based and not found machine based by selecting best rule out of 1000s of rules) not only show positive result over 1000 trades (which we all know random entries on 1min chart are always negative with transaction costs involved).

Also, the strategy is not based on an optimized variable, but which takes advantage of intraday noise or patterns (or whatever any one wants to call it).

Then if this testing was done and signal was only created on NQ then how do you explain that it works on the other futures markets i posted about earlier. Surely if it was random walk and this was simply best selected optimized rules for one market - then how did exact same set of rules work on all markets and show very little correlation between the results ?.

Please bear in mind, the strategy was not optimized or even changed to work on other markets (just backtested when strat was already complete). Also, if you see the image i uploaded earlier it shows that most markets had well over 700 trades also during the 2007-2013 period.

So now i am really curious on how you can explain this random walk - mathematically i would say it seems extremely improbable that positive result after transaction cost is included is possible.

Again i am not disputing you statement just would like you to fill me in on this particular case. I am not a math wiz or have deep understanding of stats like some of you do - i am simply a discretionary trader that could not stick to his plan so now has gone the automated route.

 

[Dhalsim]

Quote from Swan Noir:

I'm in total agreement with dbphoenix. What the hell do you think PA is? I just don't get guys that need to break it down to some simple equation that they believe is backtestable for a million years.

Context, including (in a major way) S/R, is KING. I have a really sweat setup that is currently treating me very well in CL. But I don't even look for the setup until I see either the bulls or bears strongly in control and I know I want in. The setup is, for me, only a mechanism that gets me into a trend at a sweat spot more often than not. I'm loooking for minimum MAE and a good entry bar. Does it mean the trend will continue? Certainly not but it means that when my read on context is right I have a nice positive expectation. That's what I seek.

I always thought price action was trading non indicator based and using H/L/O/C data or any S/R, Trend-line, Pattern or candlestick etc data on any time frame or bar.

Okay now i understand what you are saying but can you tell me how do you quantify whether bulls/bears are in charge? Is this a gut feeling of what you see? Is this based on any hard rules?

Also, you mention getting in on a trend, how do you define a trend - is it HH and HL? Also, how would you know your entry is not just luck? Have you backtested it or are you just trading it live and hoping it continues to work? Do you know if your entry has an edge?

 

[Dhalsim]

Quote from dom993:

I would say it is a miracle if your strat performs similarly on ES & SPY, NQ & QQQQ at any point in time.

There are enough ill price-prints on these ETFs (all stocks, really) that it is near impossible not to get stopped (in or out) regularly on those absurd (yet very real) prints.

I strongly suggest buying quality historical data from TickData, and focusing your efforts on futures markets.

Agreed on all points. However, i do find with most my other strats which do not really need charts to execute tend to have high very high degree of correlation between NQ-QQQ and ES-SPY. Hence, i tend to test any strat on the extended period from 1997-2007 (data which i do not have for futures) and keep that as somewhat out-sample data (i obviously never expect strat to match its futures performance but i do want it to still show an edge).

 

[a5519]

Quote from Dhalsim:


However, can you explain to me then how something (which is not indicator based and not found machine based by selecting best rule out of 1000s of rules) not only show positive result over 1000 trades (which we all know random entries on 1min chart are always negative with transaction costs involved).

What SW are you using for backtesting?
It's necessary to be sure that by backtesting entry/exit and slippage are computed correctly.

 

[Dhalsim]

Quote from a5519:

What SW are you using for backtesting?
It's necessary to be sure that by backtesting entry/exit and slippage are computed correctly.

Only limit orders used. However, i did GMST tests and still works well.

Software i use is Multicharts and Esignal. I think Multicharts is excellent software for anyone out there interested in the automated game.

So are we in agreement that maybe post 2007 we had a regime shift in the markets. Also, testing a NQ based strategy on QQQ is to big a difference when it comes to price action/pattern/intraday noise strategies to be considered relevant.

Last question - assuming transaction costs, slippage and so forth is all accurate. And assuming this is a strat based on my discretionary approach which i have coded to the best of my ability.

Then would you guys have faith and trade this strategy on the NQ??? What would be your deactivate the strategy point if all does not go as planned??

 

[dbphoenix]

Quote from Dhalsim:

I always thought price action was trading non indicator based and using H/L/O/C data or any S/R, Trend-line, Pattern or candlestick etc data on any time frame or bar.

Okay now i understand what you are saying but can you tell me how do you quantify whether bulls/bears are in charge? Is this a gut feeling of what you see? Is this based on any hard rules?

Also, you mention getting in on a trend, how do you define a trend - is it HH and HL? Also, how would you know your entry is not just luck? Have you backtested it or are you just trading it live and hoping it continues to work? Do you know if your entry has an edge?

I didn't get fluffed about it. I only questioned the title of your thread since not all PA strategies break down. Only those based on indicators, patterns, formulas, etc break down. Basic PA doesn't because it can't.

I doubt you're interested since your focus appears to be quantification. But if you have some time to kill, you could always look at the first few posts in the Son of If You Can Draw A Straight Line thread.

 

[Swan Noir]

Look at the 5 min CL chart from Thursday -- 9/19/13. There really is no point that you can possibly be compelled to get long -- NONE. You may not see a short that fits with your trading plan but from b1 to b66 the bulls case does not even come close to challanging the bear case. As they day progresses you become increasingly convinced that buying into the bull case can only happen if they show you evidence of trend and torque. At no point does that happen during the entire morning. It is your job to see that. The market is literally screaming to you that the bears control the market and that whatever tiny bits of bull strength that stick out are feeble at best and immediately smothered.

I repeat: It is your job to perceive that and to realize the path of least resistance is lower. The bears control every time segment. BTW, at this point, I don't even have a moving average on my chart. Except for the trend lines I draw as the day progresses the chart is clean. I don't want my judgement to be clouded ... I want to see what price is doing and only what price is doing. Others can benefit from more information but I have found it does nothing for me. I have no quibble with those that basically trade PA but use an indicator or two yet it doesn't seem to help me. They are still PA traders in my book.

As you approach the European close it gets more problematic for the bulls. Being long is uncomfortable (to say the least) and the chances of the bears becoming aggressive is high and getting higher by the minute. If they have some success the bids will begin to be pulled and the bears will become emboldened. You may call my desire to find a place to get short a hunch or an emotional response to what i see. Some may feel a need to see what their back testing shows in similar situations. Given that there has been no signs of exhaution (the morning trend had solid bars but nothing oversized) and no counter punch by the bulls the bears seem to sit quite comfortably. I believe the race is not always to the swift nor the battle to the strong but I also believe that is how you must bet it. There is plenty -- PLENTY -- of evidence that the bears are firmly in control. And absoloutely no evidence of bull strength -- NONE!! In 90 minutes of modest movement the bulls never made a serious attemp to buy against the trend -- again, ZERO ATTEMPT -- NONE!!

Not a single observation causes me to want to get long. I don't believe I am alone in that opinion and -- baring new information -- I like the short side a great deal and suspect those that trade more size than I do like it as well. I also do not like the long case at all and I am certain that unless new information unfolds neither do better traders than me. I had some family busiess to attend to on Thursday so I can't say these were real time observations. Most days general conditions are more balanced. On this day the bulls never exercized any control and those that have been trading more than a few months that could not see that plain as day need to reevalute. Setups only count after you know whether you see the bull or bear case clearly once you 30 minutes to an hour into the day. I define the day -- RTH -- as pit hours. In CL 9:00 AM to 2:30 PM. The open often has a different dynamic and a different rule book.

Quote from Dhalsim:

I always thought price action was trading non indicator based and using H/L/O/C data or any S/R, Trend-line, Pattern or candlestick etc data on any time frame or bar.

Okay now i understand what you are saying but can you tell me how do you quantify whether bulls/bears are in charge? Is this a gut feeling of what you see? Is this based on any hard rules?

Also, you mention getting in on a trend, how do you define a trend - is it HH and HL? Also, how would you know your entry is not just luck? Have you backtested it or are you just trading it live and hoping it continues to work? Do you know if your entry has an edge?