and there's the rub.....we spend all the energy in creating indicators that track as close to price as possible...which only begs the question...why not just use price instead?
Low Pass vs SMA
- Thread starter luckyputanski
- Start date
and there's the rub.....we spend all the energy in creating indicators that track as close to price as possible...which only begs the question...why not just use price instead?
IMHO:
People examine price by tick chart, range bars, time-based charts, etc. None is suprior to other in all market conditions. Otherwise, all traders would have used just one of them long time ago.
Price-based indicators are just different presentions of price. For example, CCI is basically normalized price. By examining price via different angles, obscure features may become prominent. In addition, some indicators may go beyond price by including volume, bid/ask size, etc.
Just my two cents.
Price moves in fits and starts, and is notoriously random, and yet not random enough in a purely Gaussian sense. In a way, all technical analysis is an attempt to separate the signal from the noise, and create a more "well behaved" data series. Of course in financial markets, it may well be that there is no noise, that it is all signal.
But, I like the intellectual pursuit of it. I'm a big boy after all, and when I lay my money down, and possibly lose it, it's nobody's fault but mine. Who knows, I may just get rich after all.
If all you're gonna do is buy and hold, not much point in hanging out in a trading forum.Quote from ronblack:
Good point! Following price is called buy and hold
In S&P 500 it has worked well since the low of 2009.
The main goal of trading is to do better than the market, e.g., Ed Thorp's book Beat the Market.
A large part of the problem, IMO, is that a large number of tradable patterns actually generate high-frequencies, making HF noise filtering more of a liability than an asset.
For example, V tops & bottoms, flag/range breakouts, gaps, etc.
Moreover, when those high-frequencies are not present, price-action looks sluggish & not attractive/conducive to trading.
I find it much more effective to define what I consider to be a signal or useful piece of information (data modeling), and look for it in the raw data (transformation). Most of my trading systems rely on identifying Highs & Lows (transformation), and from there only process that information. Market Profile would be another example of a possible transformation.
Statistical analysis is (obviously) only possible once a data model has been defined, and price transformed into that model.
I know very little on this subject with respect to trading. The only thing I know about LPF's are when using electronics. Never in a million years would i think that someone would apply it to trading. Personal feeling on the subject is that it's overkill but it is extremely fascinating so if someone would like to PM me i'd love to trade ideas.
Anyways,
It doesn't matter if it works or not over the long period. People want to go back 50 years for backtests for some reason only to disprove a level or moving average. I prefer to place more weight on what's CURRENTLY working with far shorter look-back periods.
The key is what's working at the moment WITH THE TREND and against it.
You can either plan on failure, or the level working, OR BOTH on separate timeframes.
Once price slams into resistance (assuming you're trading with the trend) the failure of what currently works is one way to gain an edge.
Anyways,
It doesn't matter if it works or not over the long period. People want to go back 50 years for backtests for some reason only to disprove a level or moving average. I prefer to place more weight on what's CURRENTLY working with far shorter look-back periods.
The key is what's working at the moment WITH THE TREND and against it.
You can either plan on failure, or the level working, OR BOTH on separate timeframes.
Once price slams into resistance (assuming you're trading with the trend) the failure of what currently works is one way to gain an edge.
We may have a semantics problem here.
To me, V tops and bottoms, price gaps, etc. are not HF events, they are singularities. The word "frequency" to me means repetition, and for example a V top isn't rapidly followed by a V bottom which is then rapidly followed by another V top, etc.
A singularity is a low frequency event IMO and therefore something you don't want to filter out. Of course all this assumes that you want to trend trade. If you want to RTM trade, it's the low frequencies you want to remove while retaining as much as possible the HF components.
Nobody honest ever promised trading would be easy.
