TF combos

line patterns



06302010
This thread records the progress of the development of 'line patterns', my future trading approach.
Post #1 is open to interpretation of any TF, being anchored in the 1' TF.
Post #2 includes the 5', and post #3 the 15 to 30'.
Post #3 should bring the inclusion of all TF's into my approach that uses line patterns.
From there, my trading approach turns practical and adaptable to any situation, whereupon anyone -- me first! -- can train in it.









on using various approaches, TF's to make bsTL's

Combining multiple-TF and -approach bsTL's is an art of using several charts at a time in a way that reasoning with what each chart presents is fast enough for trading at any speed.


________________

my view of multiple TF's

Time frame combinations, from whatever time frame you prefer to trade from, produce different reasons to make a trade and thus affect how well your trade goes. So, I wrote some variables to begin with.

-----

reason
Balance (+), deviation (</>), channel (=) -- data-view reasons order. My pattern recognition (for me, using a table, others using crossing moving averages or other without explicitly paying attention to patterns) explains how balance's different varieties can channel in areas and deviate from within a pattern (or an indicator's results). There are short channels; and, disparate angles may control a deviation.

time frame

1. just-touching, small-area of one screen
2. just-touching, medium-area
3. just-touching, large-area
4. somewhat condensed or several candles-just-touching screens
5. one candle holds several candles

time frame * reason
1a one reason, one time frame
1b one reason from one TF and a cursory view of a second
2 one reason each -- one TF, and a cursory view of a second
3 one reason, taken from and weighing two reasons: One reason with a weighted view of each of two TF's, and another reason with the opposite weighting
5 one reason, taken from and weighting three reasons: The last type of reason, which weighs two reasons, and one of the first three types of reasons

effort * time available to produce a reason
-- depends on type of run going for. Most multi-run reasons seem compound, of TF*reason of 2,3 or 5 - type, even when crossing at the same bars for different reasons. Effort may be changed due to nontrading reasons.

conditions of data
-- An easy data set, or one that produces lots of high-margin pips from a simple type-1a reason, does happen. Coincidence and going with the flow, planning with different sophistications of reason, and other activities adapt one to changing conditions of data.



-----

Part of the bsTL microverse: Overview, patterns (or, those with fewer or more candles), TF use, trading. (And its future: automatic; precise.)

Spread patterns from different sources would, with practice, make for improved lines. That's the idea for below:




Four charts:

Upper left, longer-term 5'+ -- logarithmic cycles : as in 'moving averages help draw' thread
lower left, medium-term 1' -- pattern areas : as in 'EUR/USD Patterns', #61-2
upper right, short-term 1' -- small groups : as in 'balance lines' thread
lower right, various -- selected bsTL's : other charts'; defined in 'bar channels' thread


In this version, I've broken up bsTL's into lines on either side in a severe angle (ul), above and center-like (ur), and going smoothly through several lines (ll).

Overlapping different TL methods at or around one TL is one approach to dealing with TL's. A more traditional view is to take one TL method per TF.

To me, the gbp/usd cycles best in the 15-30' TF's: You can see why the trend changes from one segment to the next, and observe enough of the overall flow of the cycle to get an idea of its path.
On the 30'-1H, the directional framework -- one organization I just thought of may be framework (up-and-down and side-to-side), anchoring framework (up-and-down and side-to-side, plus establishing diagonals), angle-based (a reaction to being anchored with another area's diagonals), or anchor-based (angled, only more about the local anchor than the diagonal) --



for which the cycle goes through is visible. On the 1-5' TF, details of each segment of the cycle's trend change are seen.
On higher TF's, the cycle is repeated: The 1W has very thin data, so detail like the 15' has works for a very few cycles the 1W can have; the 4H-D, then, act the 1-5' in this case.

Another poster, here, linked to an article on seasonal biology and its effects on mammals. I got several other links (1,2,3,4) and happened to notice something about our own bodies (1,2).
This convinced me that time is a worthy inclusion to trading, in that traders, being people, have little choice except to use time.

Earlier, zznbrm in this thread posted some great indicators displaying vertical lines at a chosen time of day and the beginning of the week. I've occasionally been testing them for some use that took very little of my time, since I need most of my brain effort devoted to drawing TL's and other regular shapes from my current organization (read: 'method') of them.

The built-in horizontal level with the bow indicator failed, a Gann Fan -- and complement of fib retrace -- with same did okay only took my time up and could be replaced with a lower TF's. So this led me to the daily lines.
Around my last test of the indicators, I tried the statistically slowest period of the day, 2-3 pm EST I think it was, and this showed promising, inconclusive results from whatever predictive temporary pattern and/or break between big moves the selection of time may provide. This test, I put in the US stock market's retail open and close times, 9:30 am and 4 pm EST, since the US currency, being the central and base currency, involves, statistically, something like 83% of trades.

I'd also noticed in the last test that drawing a line halfway between the week-open and daily US session vertical lines showed promising results toward getting big directions/moves. So, this test, I finally drew a line halfway between 4 pm and 9:30 am EST and another line halfway between 9:30 am and 4 pm EST; that is, I drew a vertical line from zznbrm's at 12:45 am and 12:45 pm EST every day.

The last selection of time resulted in an extra-good break between big moves. Nothing astounding, of course, or someone would have been using and popularizing it already. The point is, though, that the idea is sufficiently justified and reasonably accurate enough to be dependable all of the time.
I've always needed a way to divide the chart into periods, and this way seems okay.

Here's some examples, from a TF higher than my usual 1':


The start-of-week vertical line, , is much less reliable to determine big moves, since, I guess, it's much more obvious a place to look at; as noted earlier, though, Gann Fans and, I imagine, other instances of use of drawing devices would do well to relate to this vertical line:

It'd likely be better to have a circle around or dot at the start of the week price and time, rather than a vertical line; only occasionally does the vertical line call the week as a big move or at least a proper division of volatility with direction.

Hello.

I'm still on the 1' TF. Got rid of alot of lines, interpretation oft partial -- where you ask, what is this based on? -- still. So I guess below is a revision of post#1.


If it's okay, have a thought on timeframes.
It matters more what kind of and how good of an interpretation of the chart material you do, than which timeframe you use to do scalping, short-term/intraday, swing/intraweek, etc.
I usually look at the number of candles I have on the screen, rather than what timeframe I'm on, and base the type of trade I do on that.



Don't ask me why,
but I prefer to think of a chart in terms of how many bars/candles I have to study per screen. (I'm pure discretionary: PA and a few drawn lines.)
If it's 50+-25 candles, I try to be a hero and gut-read the flow of the candles, sometimes adding an errant line or two; if I get confused or stubborn, my trade reason is gone in a pip or two or it's exactly counter to a more important reason.
If it's 100 +-25 candles, I either overemphasize the formations to add lines for a fast, local trade, or hold off while revise until I cogently explain each and every move to include area in an overall picture and so get a lasting line.
If it's 200 +- 75 candles (my usual), I can find formations easily and get lucky and find a pure formation upon which the lines go easily, or I seek to figure how small trends' angles add up for a direction.
If it's 400 +- 200 candles, I can see and draw lots of compound lines (channels, repeating circles, etc.) on the overall picture, especially how different formations and small-trend angles interact with each other and where the special formations (circled), especially the ones that insert both balance and momentum to the current and/or upcoming trend.
If it's 1000 +- 250 candles, I get reductivistic, and either work with the super-big formation, usually momentum-based, I see, and draw a compound line (Gann Fan, maybe) on it, or detail down to a tiny balance trend upon which lots of candles seem to hinge, and draw a simple line (support-resistance, maybe); or, I select a smaller area and expand it for study.

Most of the posts above look to be in the last category, and I try to apply one or two compound lines to the entire chart screen. That seems rushed, now, with at least a high in-trade drawdown possible. Others of the charts have several compound and single lines on them; they appear static, in that while I got an accurate reading of the present material, I didn't find a way to extend from the chart into the next screen. I think I used this last category more since it offers a pseudo-high TF, and I needed at that time to compensate for my poorer understanding of how generally to interpret a chart. The charts show lessons I may use now.

Sorry to bother, really a note to self:

I just did something that, theoretically, gets rid of the last post's attention version of multiple time frames. I made a trade on one 1' chart area, and the trade margin-called (on demo simulator) and I quit that day. Then, I opened the next day for trading and repositioned onto this chart the trendline I used for the last trade in the last account and day. Then, I opened a trade in the new account for an amount that matched making the trendline of valid usage (see post in journal of today).
Guess what -- After another day on the new account, the trade's TP went through. And, I reused the trendline after that with a new trade, building my new reason atop the old reason.

In other words, I showed -- to myself, at least -- that it is feasible to discover a) ultra-long-working trendlines that fail for a very long time before working and b) the effective use of any trendline.
Perhaps b) is too much to ask for. a), it would seem, is not.

Thus, I don't need any timeframes, except, as noted, for boil-downs to get good areas easier. I can just find these especially-effective patterns on the 1'TF and apply them anywhere.