Price action with training wheels: Three Strikes Method

Look at the examples in the first post, I showed Sept thru Dec last year, October being a bad range month.

In my previous fib scalp, I found 310 pips by sitting 12 hours a day every day all month and taking the losses along with the gains. Now I see 300 pips in that month with 3 trades and a few minutes a day.

Another thing I like about doing things this way for example right now I just finished one trade and I don't see any setups for today's new bar (and tomorrow's activity) so I already have a long weekend - nothing to do until Sunday evening now.

Have a nice weekend!

Thexder....Hmmm, I recall a game for the Apple IIGS by that name, any relation? Great game for its time by the way.

Looking at your method...You wait for a peak to form (uptrend) or valley (downtrend). Market can't be ranging. Volatility must be present within the trend to form peaks or valleys, a smooth uptrend/downtrend is no good. Orders entered for the next bar (after peak/valley is noticeable) at the peak or valley. Stops....no idea....

Close enough. I first saw this game title in the 1985 or 6 Radio Shack catalog (which I still have in the next room on a book case, I am slow at cleaning my room) for the Tandy color computer, and I think over the years I found an IBM version of it and downloaded it. It is indeed my source of inspiration.

I use a less detailed analysis of what's happening to avoid needing discretions and doubt about what I'm interpreting. So I don't pay attention to how the peaks are coming to be, just that something has been trending clearly at some point that we can use as a starting point (such as 3 or 4 bars in a row with clear trending momentum), and from there I can say the trend is up or down and I standby for a way to get in either with the trend or against it as mechanically as possible.

So when a trend is going on in a very ideal example such as 3 bars in a row with all higher lows and higher highs, then something changes (a high or low or both are suddenly not in sync), the question I had when developing a system is "What mechanical observation can I make to decide if I can trade this disruption, something I can repeat blindly and take a few hits but overall more gains, and also how do I know when it's time to take a trade with or against that previous trend and avoid the other direction, or when would I be able to consider a break out in either direction and take whatever comes?"

Big question...but since it is purely mechanical, the answer is sitting on those charts waiting to be discovered as I did. All one has to do is stare at it and see what unfolds, especially since I've marked the actual trades - makes life easier than the blank canvas approach.

So let's hear some ideas. Eventually someone will get it or I'll cave and post an essay on it.

By the way I've done some side research on this, I developed it on Cable and I think it holds successful in the past on random month samples I looked at so I like that fact, but I tried applying this system to Eur/Jpy and it doesn't work at all - so I've really just gotten used to cable's personality and found a way to use it. I'd have to start over with any other chart (and I will try to).

I'll talk about why I'm taking fixed +100 targets (sometimes leaving a lot unclaimed after exiting) and how the stops generally work.

It's still a work in progress so for now I've chosen fixed targets rather than trying to let the profits run because no matter what I try, I might be able to lock in a long running trade (using a lower time frame for support/resistance stop trailing or whatever else) but it turns out more often than not I think if I tried to do that I would get caught with a loss or small profit (less than 100) while trying to let the trade develop.

So when I see a trade where I exit at +100 and it goes on for +500 before turning around, I don't feel bad. For the next 5 trades I might try to let them grow but they reach maybe +150, turn back, get noisy, and take me out for a grand total of +20 or -50 or something, and in no time at all I have eaten up that one run of +500 I got in on. Once I'm satisfied with the success of this approach I could probably then try to focus only at the trade setups documented and out of those, see how I can further narrow down the observation to guess on the probability of a trending or scalping type break out and handle the stops and targets accordingly.

I chose +100 because of my experience with how far Cable moves on average and just brute force observation - when one of my break outs is triggered I find I can reach +100 at least before it fails or gets too unstable. That figure is still up for debate, I might have to account for things like spread and allowing padding beyond the entry price before getting in the trade so I might end up only taking +85 or 90 in the end, I'm looking into it.

I didn't want to consider multiple lots and scaling out after the fixed target with the rest on free trade because for one thing that would start requiring more than once a day set-and-forget maintenance. I would have to be there when the trade goes in profit and move stops etc. I can't use MT4 to trade with so I can't use an EA. I'm sure my approach would be inspired in that direction if I could use MT4 but I'm stuck on a web based java platform that I have to control myself.

Initial stops are behind the bar I'm trading the break of and then I trail as new strong trending bars form. That means not only does it need to break beyond the previous bar's extreme while in the trend before I trail a stop behind it, the close price also has to be somewhere out there in the newly reached price area. If a bar makes a new high but the close is back within the pre-break zone I won't trail the stop yet because from observation this is a sign of weakness still and more retracement or consolidation might occur between the current price and my stop - there may be a series of pullback bars with counter trend highs and lows which can look scary but I've found it best to allow it to move within my chosen stop range and see what happens. If it resumes the trend and finally makes a new high and closes out there somewhere, I move the stop behind that bar finally. In the examples from September 2007 there's two trades that show stops being kept at the initial bar during a rocky road and are eventually trailed up. If you look at the close prices of the bars you can see I moved the stop when a new high was made with a close strongly biased that direction also.

Take a look at Nov 17 2006 cable daily - what is that, 800 pips upward for the rest of the month? I would have been in at the very bottom of this up move, but I'd have been out at +100. Even a trailing stop would not have been taken out until the top of this move if it were blindly moved on each bar.

This is why I want to figure out how to stay in the trades after I do more testing at the level I'm at for now. I tell you there is a mechanical way to know when to get into a trade at a turning point in the trend momentum and all you have to do is look and find the answer I found. Now I'm in the same boat with staring at it and finding the mechanical way to stay in the trend. I have no idea how right now.

There is one advantage in this specific case that could be implemented without changing the real rules and still maintaing a set and forget daily maintenance condition. There are a few days of opportunity to think after the trade is entered and before the fixed target is reached. On the day before the target is hit the stop can be trailed up so the trade is almost free. We could try keeping the stop there (now that there isn't much to lose) and letting it ride without targets.

Or test a new possible idea where if targets aren't met on the day of break out and the trade is live when being re-evaluated the next day, switch from fixed to trailing stop with no targets, and put the trail at the same amount as the original stop size would have been since this is a way to normalize the current range of price movement in a single day. If we get any trend continuation and it reverses we may get some profit but if it takes off we may get much more. In a trade like this one we could possibly come back the day after setting the trailing stop, see it's still in play, and by now we see it's very much a smooth up trend so we lose the tail, go back to fixed stops behind the previous bars and handle it manually day by day (by now it's free trade minimum).

That would still miss some of the "big" moves that occur on the same bar as the break out but it would allow some occasional gifts like this one where we get to intercept. Of course along with testing to verify that this doesn't screw up a bunch of other trades as-is, we would need some conditions to be met like the bars should be trending strongly during the break out (new high and closing out in the new high area). We wouldn't just switch to trailing stop for any trade that stays in play overnight.

There has to be a way to get those 800 pips. But in other cases where 500 pip moves occur, the first target is hit right away after break out so there is no time to intercept. There still must be a way to determine when to let it run free.

Another situation where a trade is taken and target isn't hit immediately so there is time to observe a very strong looking close on the last bar, an overall trend-looking time period, maybe worth implementing the alternate stop strategy. In this case the fixed 100 pips would become 200 pips.

I decided to try a noisy month to see if things can go very wrong. Feb 2007 was all chaotic. I saw 4 trades I'd have taken, two good and one bad balancing out to net zero for the month (better than a loss). I'm still content. I have seen other months like this with a break even or small profit/loss followed by a month of good returns to make up for it.

...again, i thought i might have been getting closer, but then looking at this chart i notice you don't always have to have a strong close to place an order?

Some clues:

1. In my approach I only ever say the trend is up or down. It is never "consolidating". It might be "pulling back" and may linger in that state for a while but I always have a defined trend. I toggle the trend direction back and forth based on what occurs with trades. When I take a trade in a direction, I set the trend that way. If I trade the opposite way the trend reverses for me.

2. When I am looking for a long trade I only look at the tops of bars. When I am looking for a short trade I only look at the bottoms of bars when making choices.

3. Depending if an upcoming possible long or short trade is with or against the latest defined trend, the evaluation of those bar tops/bottoms will vary. In other words, looking for longs I look at the top of the bars - if the long trade is with an up trend I'll evaluate those bar tops differently than if I'm trying to go long against a down trend.

There are some glitches I'm ironing out yet such as whether to toggle the trend direction in certain cases as soon as an existing trade is done, or wait for something to happen. But it hasn't impacted my test results so it's just minor.

Correct. I consider strong closes regarding stop management.

It's probably easiest to look at ideal trending charts to develop theories and then check them against the noisy examples to be sure.

It also might be easiest to start with trying to figure out how the trades with the trend work rather than the counter trend trades.

are the entry's based on the bars having matching lows or highs, or double bottoms, if ur using price action i guess.. Its hard to see the charts as far as what the bars look like but let me know if i'm close

I'm not looking at the traditional patterns in this case, so no DBLHC, IB, TBH, etc. It's just totally made up from observation brute force. I consider it price action since I am looking at what the bars are doing and no other indicators of any kind, not even moving averages (simple 1 period are on the chart but they are the same as the bars themselves connecting high, low, and close prices for a visual impact - the training wheels).

Maybe if I tried to learn the traditional method it WOULD be universal...

January this year.
Regarding the stops again, in some trades it may look like it would have been stopped out before hitting what I write as a successful +100 like the first long in this month's example, but the stop management is based on what happened before the trade. So in the long trade here I have the Buy circled and the following 2 bars have lower lows, but that first low was established before the break out happened, then end of day came and I saw I was in a trade but my stop should really be below that new low, not my original estimate higher up when I set the pending order. So I shift it down and it doesn't get hit, targets are met.

And for those that are more familiar with my tricks, I should also point out I'm not cheating by going to lower time frames for more info. Everything is done from just the daily chart.

...i've based the chart observation on simple HH/HL for longs and LH/LL for shorts for April so far (ODL feed), Sunday bars included, which i don't think you include...

The trades are similar to yours, but i have a loser...just wondering if i'm getting any warmer?

Getting warmer - here's some pointers...

Remember my list of clues, I only care about one side of the bar at a time depending if I want to go long or short, so my idea of bullish wouldn't require the full HH and HL criteria etc.

Did you choose your bullish/bearish bars based only on how the bar itself appeared in the context of the chart (having the HH/HL LH/LL) or were those chosen for other reasons (my clues about switching my trend state from up to down based on trades taken, and what is happening on the potential trade side of the bars are more relevant than if a single bar is bullish or bearish in appearance. I'm thinking more about overall momentum in a sense rather than looking for single bars)

That losing short almost got me too when I was evaluating the setup but I decided it wasn't time to go short yet despite any overall bearish sentiment implied by the chart. If a break had occurred sooner it would have been a valid short for me but it took too long creeping downward bar after bar so I disqualified it.

I'm also wondering about different data feeds and how they might impact the whole thing. On my other system I use a different demo feed that shows sunday bars and my main feed doesn't, and the time zone is different on the two charts so the data is a bit weird. I'll go over some of the same trade setups on that data to see what happens. It might be another case of "trade what you see" and so I might have totally different trade setups but hopefully it all balances out. Maybe I even get some good ones with one data feed that I miss on the other.

Don't forget to use the training wheels (pictorially). I wanted to just be able to look at the chart from a distance as a picture and see what is going on (at a distance I can't observe much detail about what each single bar is doing). So I connected the tops, bottoms, and closing prices and zoomed out to observe.

"When a trend continues, what seems to occur on the chart most often at that time?"

"When a trend reverses, what seems to occur most often?"

"When a continuation or reversal is a failure/fake out, what has occurred differently from the successful times?"

"Hmm...I wonder if this works more often or if this is a coincidence"

...one + years of back testing later....

"I think this might be it..."


That's why this system is dedicated to Cable, it is based on observing that personality and how it moves. When I tried it on Eur/Jpy the behavior was nothing like Cable when the good moves came, but I started to ask the same questions and I started to find answers because I had learned how to "read the chart" and interpret it. So I'm working on a Eur/Jpy daily system now too.

The latter, take trade 1 in the chart for instance, the prevailing trend was down, so looking for short breaks of the previous bar's low...the bullish reversal bar printed instead suggesting trend change, so looking for a break long.

There's clearly more analysis to this than meets the eye, and the discretionary element is perhaps what adds the 'fizz'...will persevere

At the beginning of the month when the prevailing trend was down from the end of the other month, suppose that bar wasn't bullish - what would be your course of action if it had made a lower low instead of higher low? Or any other possible type of formation, inside bar, etc.

The discretions involved are not really a major element in the trade setups. In the case of that failed short that I say I avoided, that was from a kind of discretion but it was still a set rule such as "If it takes too many bars to break out...cancel it" or "if it is too far away from a break out level, it's too risky, cancel it". That kind of thing. But it's "fixed". That's just a special case though. Those other 3 long trades were completely mechanical.

...if it made a lower low, i'd probably be wanting to trade a break of that low within the next few bars..