Price action with training wheels: Three Strikes Method

Discussion Refocused

Refer to the new thread http://www.forexfactory.com/showthread.php?t=82180. Since the start of this thread the system has evolved in one month to a whole new realm that required its own separate mentality so I have moved the thread.

Live forward testing has started this week after a month of back testing and tweaking of guidelines. This is the chart for the Sunday/Monday session this week.

The peak indicated with the blue arrow is the only possible trade setup I would have even considered but just by the volatility I disqualify it from my experience with Cable. This widely ranging parallel bar action includes volatile bars running in the counter-break out trend direction shortly before an attempted break out at the blue arrow, only to reverse suddenly and move another great distance in a short time. The risk of whipsaws is great.
The initial stop would have to be too large also.

This method involves sitting on the sidelines waiting for more calm setups. Today would have been a rewarding day for a longer term method.

That first possible peak at the left of the chart isn't a tradable peak by the definition of it being an irregular noise movement of the upper wall. During the time of that hump formation the tunnel walls are either diverging or converging as the upper wall rises to that high, so the tunnel is considered to be moving sideways or continuing downward from before rather than moving up to form a proper peak. Better setups from clean price action are desired even if a break out of that upper wall peak should prove successful. I aim to try to only take the most confident setups for now.

The low peak near the end of the session is outside normal trading hours and the tunnel is still too irregular, and seems to be trying to calm down, consolidate and come back under control. There is a great possibility that it will just move in small waves for a while through the off peak hours rather than going for yet another big down move so soon.

The first day of the week is avoided.

This is a good first trade to kick off the first week of forward testing Three Strikes.

Took this long trade for +161. It moved straight to the top with higher lows so a manual trailing stop preserved things well above 100 pips locked in profit, then I made an educated guess and put a trailing stop so I could let it run on auto pilot. I did a little better that way as it started to collapse, getting possibly 30 more with the trailing stop than a manual stop. I chose the value of the trailing stop by observing how much retracement had occurred during the up trend, then set an automatic trail somewhere slightly outside that range to allow the controlled up trend to progress as it had been for hours. Since there was a sudden up surge before the collapse, the stop trailed up and caught the extra pips. That's not part of the strategy but I was baby sitting the trade all night and needed to go out since it was morning.

Initial entry was 2.0054 including spread and initial stop was 1.9996 including spread, below the pre-break out bar.

When it went into +50 profit I moved the initial stop up to break even since I've lowered my tolerance for seeing profit turn to loss. I'd rather see profit turn to break even.

When evaluating the possible long break out here's what I saw:

Strike 1: Bottom tunnel wall is facing counter trend
Strike 2: Price line is mostly short biased inside the tunnel

The only other possible issue is with the divergence of the tunnel walls the tunnel technically got wider but it's only minor and if it is counted, it's still only 2.5 strikes. There's no other issues.

If that tunnel width issue was to become a go/no go factor I would probably allow the trade because the tunnel is so narrow that a little broadening is expected at some point without cause for alarm - the filter is meant to detect sudden out of control ranging of the bars and this doesn't appear to exhibit that.

The tunnel (price bar movement) is well behaved. Another explanation to justify the broadening is the trade is finally winding up to break out - it is very likely going to want to move a little counter trend and then blast off similar to the way the first break out of a coil range is often false as price winds up for the true break.

In the beginning of this session not much was taking shape prior to London. The red lines on the left show areas I wouldn't consider tradable peaks. The upper one is just a consolidation getting very tight and not very much of a peak. The lower one is a pull back downward but then a jagged tunnel wall blip goes slightly lower than the actual tunnel pull back low, so it's also just a consolidation area. I would want to wait for something to start taking shape and preferably form brand new peaks to trade once the momentum is building.

The first two tradable peaks I saw were the two shorts marked with an X.

The first one I didn't trade because
Strike 1: The price line in the tunnel was favoring the long bias
Strike 2: The tunnel walls were heading upward counter trend at the time of a pending break out
Strike 3: The previous daily close was inside the bar range from two days ago, risk of ranging
Strike 3.5 and 4.0: The tunnel was irregular looking still and getting wider - too much risk of unpredictable trend direction or ranging.

The second low peak I didn't trade because
Strike 1: The price line was still not short biased
Strike 2: The previous daily close was inside the bar range from two days ago, risk of ranging
Strike 2.5: The tunnel just got even wider still

The trade setup was starting to improve but I noticed the last two bars that caused the tunnel to widen were continually increasing in volatility, and at this point the initial stop would have to be around 100 pips. Sometimes that much of a stop would be considered normal for the setup but until the volatility stabilizes I don't want to take the chance that the next bar is even bigger and takes out a stop like that.

The 3rd short setup was traded. The volatility came under control plus there was a pull back and the tunnel started taking a parallel form. The price line in the tunnel was definitely short biased, hugging the lower tunnel wall. All irregularities in the tunnel appearance - wall shape and width, were normalized. The tunnel was well below the trend line unlike the early part of the session where it was starting to envelope it. Out of the recent bars formed during the down leg of the last wave and the pull back bar, the distance from the high of the break out bar down to the break out price would be smaller than the most volatile bar that's formed recently so although the bars are all large, it seems like a worthwhile setup.

Strike 1: The tunnel was aiming up prior to break out
Strike 2: The previous daily close was inside the bar range from two days ago, risk of ranging

Trailing the stop on the hourly bars gave +65 by the time the stop occurred.

In this session there's no tradable setups. Also I'm only going to consider setups during the London and NY time periods for now.

The first London setup is a short that gets disqualified because

Strike 1: The trend line is trapped in the tunnel for 4 bars before break out
Strike 2: The price line has been long or neutral biased before break out
Strike 2.5 and 3.0: The tunnel has just widened prior to break out and is looking a little irregular. During the Asian session there was a hitch in the lower wall and leading up to this break out the top wall is getting a little square while the lower wall is rounding over. Something is strange and the overall sentiment is consolidation with a counter trend bar closing bias.

The potential long trade has a lot of debate about it. In its favor it is a horizontal resistance that's had 5 prior attempts at it over two different attack sessions so breaking this area could lead to a big move. But there's also been considerable ranging with inner battles that haven't yet been settled. There's been a recent highly volatile bar spanning the entire session range without being able to break the top, more tunnel irregularities and no distinct wave peaks in the tunnel movements. The trend line has flattened out and everything points even more strongly toward continued ranging and a fight within the range that has no signs of clarity for an emerging trend.

Strike 1: The trend line has been engulfed in the tunnel for a long time
Strike 2 & 3: There was a recent volatile bar closing counter trend
Strike 3.5: The tunnel continues to be irregular shaped and going sideways

Unless I see something very strong take form, I think I'm taking today off.
There were no setups yesterday and today is a holiday. Yesterday closed within the price range from two days ago giving a risk of consolidation and it's been doing just that. At the moment the price action is extremely narrow and riding along the flat trend line. The tunnel has been highly irregular for a day and something has to start forming better waves at this point to make any peaks worth trading. It appears to be just random for now.

Here is a summary of this week's trade signals. Out of the 5 days only 2 trades were signaled after going through all the constraints on the hourly chart.

The system was successful in preventing trades during times of indecision and also successful in detecting the start of two trending moves and getting in on them when they proved their stability.

whether or not this was just a lucky week with ideal setups and easy to avoid ranges regardless of the system, the point is once the training wheels eventually come off, all one needs to do is zoom out on a chart over the past few sessions and be able to clearly see an emerging trend and know when to take it.

The last thing I want is to have a system that is tedious, but everything is awkward at first when you're learning it.

Now that I've revised the setup criteria several times since the first back test and I've also become more familiar with some good and bad types of price action, I'll try a longer period of testing and see if there's anything more I can account for in the checklist or if there's any ways to help catch missed opportunities. If I can find a way to even simplify the whole process by staying out of the market until things start looking better, I will consider things like that rather than always checking setups with brute force every bar. When you look at a zoomed out chart after the fact it's easy to see the good and bad trading times. I am wondering if I can catch either or both of these situations as they occur.

I'm working on a new revision that should make it much easier to find tradable peaks in all the jaggedness of pre-trending moves. The holiday obligations are slowing down the testing so I may not have something for a few days into the week but it involves waiting for a sudden volatile sign of life from dormancy (a sudden large bar breaking from a consolidation) and then doing an evaluation. A sudden bar like that is easy to spot so that helps identify the trading opportunity.

We're used to seeing a big move suddenly occur and then we think we missed it only to watch 3 more equally big bars follow in a trend. I'm attempting to tap into that. It should help increase the good moves captured and minimize the bad ones and make the overall method easier to follow. It's something I've been observing lately while looking at fib scalps.

The more I develop this system and trade it, the more the training wheels are already coming off and I think I've also possibly eliminated the need to evalute a big checklist. I may need to rename the whole thing soon (too bad the thread title can't be updated).

If I come up with something solid (back tested and forward proven) and it's evolved into something totally new, I'll just start a new thread. The ideas in this one are still useful for inspirations, afterall that's how I got where I am.

Still no tradable peaks since the consolidation started. The last several days have also closed within the prior day range and it's now moving in a trending consolidation...moving up finally but still going sideways so no peaks to work with. All the bars are running side by side and when a new high is made, more parallel bars form, no solid trend action. Today might finally close outside the range of the last several days and something should trend.

Still working on some documentation and testing.

nice works, go on
bj

There were still no trades in this method for the past 2 days. A pending long today at the major peak near session start wouldn't be taken because the trend line was engulfed, price line was low, and the tunnel had gotten wide and irregular. The long break was just a big spike that retraced and would have been a bad move so it was good it was avoided.

The pending short after that would have been at a time where the entire tunnel was whipsawing - both walls up, down, up (during spike up), then a break short, so 3 strikes for the tunnel whipsaw prevent a short. Now it's just moving in what could turn out to be a long term wide tunnel range.

Yesterday things moved up but without any trading peaks, just a north-east tunnel.

This method of course is designed with this in mind, just waiting for the strong looking moves and ignoring the rest. It could be argued that the two good trades last week were good enough to cover all of this month.

I'm still working on the "new" method, but it's now a background task. I'm working on 3 trading systems all at once so I just dedicate my focus wherever I'm most inspired and moving forward at any given time. But what I can say is I'm working on adapting this method to a 4 hour chart, not sure if it'll pan out. It would make it easier to work with and monitor for setups less often.

+40 pips from a long yesterday. Again this is a conservative method. The high peak that was traded is the only proper looking peak out of all the highs and lows yesterday. The circled peaks are jagged single wall peaks while the tunnel moved in just a single direction all along without really peaking, as noted by looking at what the opposite wall is doing while the circled side is peaking. If a lower wall forms a spike down and pulls up to form a low peak but the upper wall is just moving down or sideways, it's not a tradable peak.

The traded peak in this session is particularly explicit because both tunnel walls are moving up in parallel, then both down in parallel although the tunnel momentum doesn't have to be that clear to be tradable.

Before breaking long, the price line had been biased long near the top of the tunnel and then just pulled back for one bar so it was still long biased. Both tunnel walls were aiming upward so that was two strong items. The trend line was also moving up well. The tunnel was a little wide but it had narrowed prior to break out and was potentially coming into focus so no strikes were given overall for the risk of whipsaws with a wide tunnel - there were enough positives indicating an upward trend with just a consolidation waiting to break out rather than a totally undefined and risky trade time.
There were no recent counter trend volatile bars to cause a contradiction of signals either. Everything seemed to be coming under control.

It looks like there may be a coil forming between the previous and current day but there is no day to day ranging yet with daily closes compared to previous day price ranges. The interday coil can be argued because it isn't perfectly smooth with the irregular tunnel in the current session. The bars would have to be examined from one to the next and see if the tunnel had actually wiggled out of a coil pattern even if just for one bar, and it would have to be debated. Again the overall sentiment has been established as long biased or else inconclusive signals - no solid warnings of a down trend or a whipsaw to worry about.

Strike 1: The trend line was stuck in the tunnel too long
Strike 1.5: The tunnel was kind of jagged

There were no peaks to trade the ride back down after the long was stopped.

If anyone is curious what's keeping me so busy from coming up with the new revision to this system, I've been working on this: http://www.forexfactory.com/showthread.php?t=76403

I'll get back to 3 strikes eventually.

I posted a new document fragment to give an idea of where I'm heading with this now. It's still being developed and definitely needs to be tested more but I'm aiming to make it much easier and hopefully universal to charts and time frames. So right now this is all a research exercise in progress until I revise something. I'm doing a lot with scalping right now so time is limited.

I've been away but I haven't been messing around or letting this stuff slide.

Take a look at the first post (as it is at this time of writing at least, it is always changing) and see if it looks like I'm on the right track and if it might interest you. I've migrated now to the daily chart and have gotten to a point that approximates my personal grail quest - freedom in 2 ways.

1. High probability (subjective) of success and minimal or no discretion to interfere with decisions.

2. Set and forget once per day at the start of the new bar for however long it takes to make the decision by looking at the previous day/days.

No staying up all day/night tied to the machine.
No wondering if this really is or isn't the right time to place a trade or if the trend is questionable or if there is anything else to worry about.

So far I've tested June 2007 to present and every month has had a positive balance, sometimes +55, sometimes +600 for the month. Typical return of at least +200 to +300, sometimes a string of +400 to +500 months.

That's just using the cable daily chart and as I said only looking at the chart for a few minutes when the new day begins. I haven't looked to see if the same logic applies to other daily charts. If it does, all the better, spend an extra few minutes a day and generate more results.

I am still evaluating the setups and so there is no document. I will do as much testing as possible before I consider myself on to something and ideally I'd like to test other charts too before I complete the effort. So this is a preview of where things are. It's been a month since I started the thread and I think (despite the lack of posts) it's come a very long way.

Just exited a trade with fixed target. This would be the 3rd trade for the month for +300 total without even thinking about the setups. I blindly set this order at 5pm EST yesterday and woke up to a profit. Compare the simplicity of that to the previous examples with all the decisions involved and risks of getting it wrong, and all the screen time watching and waiting, trailing stops all to try squeezing a little more out.

My initial stop for this trade was over 200 pips behind the previous bar so some may not like the risk/reward of a fixed target but that's what a year of testing (with more to come) is for. I figure I see up to 90% successful trades so far and especially with money management, even if that performance slides over time it's a winning strategy to me.

Look back at post 7 on March 21 where I posted a weekly summary of 2 big trades totaling around +200 pips.

Look at all the hard work, screen time, stop trailing.

Here's the way I view March now. I've circled the equivalent time frame from the previous post and the current approach. Same trade results in those few days but taken slightly differently and much easier.

Still no indicators, just lines connecting the dots between bars to visualize the momentum better.

+600 pips for March 2008 sitting for 10 minutes a day at 5pm EST to set and forget.

Here's the other 2 trades for this month included with today's.
I will attempt to forward test this rather than trying to define a set of rules strictly in advance.

Any ideas on what I'm doing?

lol, trying to decipher which line's which - looks intriguing and no doubt an excellent KISS approach. I'm glad you're finding/found your personal grail

Good point. I noticed some residual superfluous MAs on there. Ultimately no lines are needed except the price bars. The three main lines I have charted there are the SMA 1 high/low/close which only connect the dots between bars really, so that's why they aren't needed. The other slow MA can be deleted.

..ah OK. Thought i had what you were doing, but on closer inspection....i don't. Looking forward to seeing how your forward test progresses, particularly within narrower ranging periods.