Hi all, It's often been said here and elsewhere that charts don't contain all relevant information and that there are other ways to perceive historical data. I have some ideas about what this means, but I'm curious what else I might be missing. For instance, data can be put in tables. One could calculate statistics. Rule based strategies might be utilized, i.e., opening range breakouts as an example. Might tick data be better to analyze than say 30-second or 1-minute OHLC data? Could Excel or other spreadsheets possibly be used to create custom charts or displays of data? What are quants looking at in their models? Any thoughts on this?
-Distance above/below MA in % -Correlation -Bar patterns Anything you can see on a chart can be put into numerical format, and tested in any way you want. Other things that can't be seen on a chart too. I personally use 15 minute data because it suits my needs just fine. It all depends on what YOU want to accomplish. You need to have an idea or a need for a certain view of the market. It is then up to you to figure out how to manipulate data to get what you need.
Great topic. Charts are useful within a rule based, statistical/data driven approach, as it allows us to visualize data. What you do with that data is what separates the wheat from the chaff. Chart pattern recognition IMO is a complete joke.
People are sometimes too fixated on the petty minute technicals of a chart -- they should instead be considering the overall greater picture of emotions mixed in with the combination of facts that can or will make a chart dance or move, Expectations unfolded, or unfolding, -- Or of course, the Opposite can be happening too...and you have to be ready for it, and open-minded,
Chart pattern recognition for short term trading possibly is a joke, for longer terms eg several days holdings to weeks to months & years isn't a joke imo. I've attempted intraday trading via charts and can make no sense of it, I could never trade it profitably.
I imagine someday this will be more available. Watching the days morphing of market activity would be very cool indeed. https://quant.stackexchange.com/questions/549/innovative-ways-of-visualizing-financial-data It could be extended to all exchanges. The overview would look like a digital sunrise as a wavefront washing over the globe. Is there any promising finviz tech developing into the augmented/virtual reality space? Slippery slope into a dystopian future. I like my sunny window and garden views.
If you can mine, manage, and analyze the data properly within a strict framework, you can navigate through the noise once the stats look convincing.
Using option chains and probability cone -- both of which are free on broker platforms will obviate the need for any charts. In fact, this is much more advantageous a method than using charts. Options predict how future direction and how price will evolve. What works is not to have lagging data which is true of most indicators and instead build on probabilities which are predictable by review of options chain data.
Charts are after all the visualization of numbers. That's all. But I don't think everything is easily backtested. How can you for instance objectively define a trading range and test a breakout of it? I'm probably just slow, but I haven't yet figured out how to do this 'mechanically'. I asked the question here. But I digress a little... Thank you for your contribution. Interesting tables. Mind you sharing what the 'threeup', 'threedown', 'inside' and 'outside' means? Or is it proprietary. I do have a hunch what it means, but I'll stay silent in case you don't want to share it with everyone...