Chart Patterns

[tech/a]

Hmmm

I'm afraid I've never seen any harmonic trading methods that can be used for profitable trading going forward on the right side of the page.
You can make them "fit" in hind site but using them to trade --- You'll be the first.
To code them even with "C " or python with so much subjectivity again highly unlikely
If there was enough data I guess with advanced coding knowledge you could instigate machine learning as we have in some of our methods. (still a work in progress) but very interesting to say the least.
Well over my pay grade---that's why I pay them!

 

[barney]

I suspect it's possible that harmonic Forex "patterns" might be a bit more reliable/repeatable but that is only based on my own eyeballing.

Again just my own opinion, but each "instrument" will likely have its own "regular patterns" (whether Stock, Forex or Futs) … and perhaps optimisation can be concentrated on each given instrument?

It's very doubtful that one size fits all so to speak … but history tends to repeat when we are dealing with human behaviour (trading or whatever).

I can't even code my mobile phone so totally out of my league but an interesting topic.

 

[frugal.rock]

Great topic, love it. Will have to read through the whole thread one day.
My .
Charting, a picture of a dataset, simplified.
Useful as an indicator? Yes, to me it is. As far as patterns go, I only use general trend lines, from intraday trends through to 10 year trends, price against volume traded.
A pattern of an average. As far as other patterns, (head and shoulders, batwings, crouching tigers... is there any limitations to creativity?!) they were all formed by a particular set of circumstances. The trick, as I see it, is the ability to quantify each of these circumstances (variables).
These variables, I consider to be virtually unlimited. Examples: trump tweets, weather/climate, trading volume, sentiment, outlook, liquidity, profitability, potential profitability etc etc.
I have been working on a 'system' to categorise some variables into mains and subs as well as trying to quantify them (ie; a value or a value index with a weighting to be applied for each variable, to come up with something useful?).
Having a bit of trouble quantifying sentiment to name one of a few,
any ideas anyone?
Basic excel spreadsheets for me!
My idea overall is to use historical day trade data, analysing said data against the variables ultimately to give an outcome of signal to buy, hold or sell but broken down into degrees /range, of high and low of signals.
Will get there one day, a work in long progress! If ever finished, will then ask someone to code it for me.
F.Rock

 

[barney]

Having a bit of trouble quantifying sentiment to name one of a few, any ideas anyone?

It can be ambiguous that's for sure …. The earlier you try and call sentiment, the more likely you will get it wrong as well of course.

I quite often post up "early" calls on Stocks , mainly because making a call publicly opens you up to scrutiny if the call proves incorrect ….. and scrutiny tends to heighten your attention to detail, so its a good way to learn for me personally.

On the topic of quantifying sentiment …. The relationship of Volume to Price over Time is likely the Holy Grail of trading if you can isolate recurring patterns with a high strike rate.

There are obviously no 100% strike rates to be found, but that's where money management comes in to limit losses.

Interestingly enough, if a pattern with a high strike rate fails … the ensuing pattern following the failure often becomes a high strike rate pattern in the opposite direction for X amount of time

If the above made any sense, then like me, you are likely doomed to a life of frustration looking for the elusive Holy Grail