Too Many Indicators?

OK, here's a free lesson from a long-time trader and ex-RIA who has taken >$100 million from the markets for clients over the years. (I know, $100 Million ain't much when compared to the big boys... but I'm not on Wall Street... I live in a one-horse town near the Rockies... not exactly a hot-bed of finance and money management.)

I've used MetaStock software for >25 years. I once attended an "upgrade your software" seminar where the guy running the show (could tell he obviously was a long-time trader himself) said, "if you must use indicators, ONLY ONE!" (Imagine that. Selling MetaStock with its 50-ish pre-programmed indicators + custom section to allow traders to create their own parameters".. and the guy running the seminar says, "pick one".)

I concur. When I first started out, I bought an expensive satellite real-time system which cost $10K... thinking, "I'll gather all the info and distill it down to the market's essence and make a killing".

Over time, I came to...

1. Trade Price!

2. If you must use indicators, ONLY ONE. Which one? Probably doesn't matter. The more indicators you use, the greater the potential for conflict which results in indecision.

3. If you're going to use an indicator, I'd suggest a "range indicator"... RSI, Stochastic, CCI, or Williams %R. Doesn't matter which. Just study them and pick one you like. Correlate it to price. Then, research how it behaves in uptrend, downtrend, and chop markets... and trade with stops.

4. Trade "trendlines"... support and resitance hold/break.

You'll be ahead of 90+% of the pack.

You're welcome!!
 
CoolHandLuke said:
Raw data, and raw data alone is price itself. You chose candles/bars of a specific TF of your liking then draw lines connecting the bottoms and tops of those candles/bars of your liking. That's mathematical manipulation of the data, or said another way an "indicator".

Why don't you connect your lines to the tick price, rather than 5 mins or whatever of captured ticks??
More...

Drawing a trend line along the bottoms/tops of price waves requires no mathematical manipulation whatsoever.

But if you disagree, that's fine. This has been debated many times and anyone who's interested can do a search.
 
Unless you flipping a coin, everything is an indicator, and no indicator can predict the future. Doctors can't tell if your heart will beat 2 seconds from now. Anyone who says their indicator can predict is lying cause one loss will make that indicator untrue. As traders we use what works for us after watching the screen thousands of hours, we understand what the market should do and often know why the market does what it does so you can expect what might happen cause of times of the day. The old phrases like "Buy the Rumor and sell the fact" didn't come out for the heck of it, but after traders found it worked more times than didn't work. And that what indicators do whether moving averages, trendlines or Price itself, we form rules of what to do when something fits our rules. You spend so much time with markets, and after awhile you just "know" where the greater percentage should happen for a trade and you just know where the trade is going to fail.

And in the end, we spend thousands of posts of defending something that doesn't mean crap. You either know how to trade or you burn accounts.
 
SplawnDarts said:
I use three custom indicators - one counts things on the tape, one does some linear regression math, and one is a more statistically rigorous take on Bollinger bands.

Bollinger band (or my custom take) are modestly interesting to look at, but I've never seen any evidence in testing that they have any predictive value - I mostly keep them to autoscale my charts.

I've never had .....with trendlines or MACD. MFI and accumulation/distribution are OK ideas I guess, but pretty stupid in the modern era since you have the tape handy and can just count up transactions.
More...
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

I've put bollinger bands on volume, but i dont think they were designed for that;
but a 20 day ma on volume can help, but thats not a prediction. Matter of fact i have caught sellers selling + buyers buying ,, just barely /carefully below 20 day ma..LOL:D But that really had nothing to do with bollinger bands; that simply the 20 day ma.:cool:

I dont really like MACD;
@ least a 50 day moving average under lines price or over lines price.MACD is a dis -traction to me. If some can use it, fine.Its for sure not a predictor.:cool: Years of screen time maybe more helpful than any indicator; not a prediction.
Wisdom is profitable to direct.
 
Scat, if i can ask, Did you ever/frequently muddy the waters NOT trading price.
An example might be that treasuries are frustrating many market participantsthis year.
"They are behaving strange" . Did this way of thinking ever screw you up?
 
Scataphagos said:
OK, here's a free lesson from a long-time trader and ex-RIA who has taken >$100 million from the markets for clients over the years. (I know, $100 Million ain't much when compared to the big boys... but I'm not on Wall Street... I live in a one-horse town near the Rockies... not exactly a hot-bed of finance and money management.)

I've used MetaStock software for >25 years. I once attended an "upgrade your software" seminar where the guy running the show (could tell he obviously was a long-time trader himself) said, "if you must use indicators, ONLY ONE!" (Imagine that. Selling MetaStock with its 50-ish pre-programmed indicators + custom section to allow traders to create their own parameters".. and the guy running the seminar says, "pick one".)

I concur. When I first started out, I bought an expensive satellite real-time system which cost $10K... thinking, "I'll gather all the info and distill it down to the market's essence and make a killing".

Over time, I came to...

1. Trade Price!

2. If you must use indicators, ONLY ONE. Which one? Probably doesn't matter. The more indicators you use, the greater the potential for conflict which results in indecision.

3. If you're going to use an indicator, I'd suggest a "range indicator"... RSI, Stochastic, CCI, or Williams %R. Doesn't matter which. Just study them and pick one you like. Correlate it to price. Then, research how it behaves in uptrend, downtrend, and chop markets... and trade with stops.

4. Trade "trendlines"... support and resitance hold/break.

You'll be ahead of 90+% of the pack.

You're welcome!!
More...
%%%%%%%%%%%%%%%%%%%%%%%%%%%%

Good thoughtful points, scata.
Not that I would want to just use one moving average, even William O Neill used 3 moving averages[50dma 0n volume , + 50+ 200 day moving averages on price And IBD uses more than that on their index charts .

Also not to nitpick or take away from your mostly good points;
even price grids are quite helpful, price grids maybe another indicator ?? Or one could simply call price grids '' price'', :D LOL
 
I do resort back to stochastics as "do it now" to take emotion out of buy sell.
Does it tell me more than price itsellf ? No, but i need it.
All else i put below a chart is a cutom indicator that measures and displays the multi day and multi week swing highs for longs and swing lows for shorts.
It can help keep me out of trouble. It doesnt make sense to trade short when i
can see that short trades are few in number and have produced little percentage gains
Vs. long trades. (Trading with the trend).
I can also scan, filter, backtest and characterize stocks and etfs with this one indicator. You could construct it fom any price container. I use the 15 bar high and 15 bar low as my container and calculate percentages relative to these limits.
I dont know that additional indicators would help but i do need these.
 
Here is why indicators don't work consistently. Because they are all based off of data from the past, and can not predict the future with anything better than 50% precision. Why? Because price is pretty darn near a random walk process.

Price distributions are not perfectly Gaussian. The skew and kurtosis you find in financial data distributions means that there is much more complex processes going on than anyone has ever been able to fully model (including Noble prize winners.) Too complex for simple indicators to consistently model.

Of course the skew and kurtosis can manifest itself in things like price behaviors around obvious support and resistance levels. Sort of a self-fulfilling crowd behavior. Ever notice how price often times blows through S/R levels, or can turn on a dime at these price levels?
 
dbphoenix said:
Drawing a trend line along the bottoms/tops of price waves requires no mathematical manipulation whatsoever.

But if you disagree, that's fine. This has been debated many times and anyone who's interested can do a search.
More...

To me you're lost in semantics. You draw an upward pointing strait line connecting the bottom of a few 5 min candles. That line forms arbitrary acute angle with the horizon the magnitude dependent of the steepness of of the upward trend. You then extrapolate that line to determine where to buy. The price your strait line instructs you to buy is simply a rough RATIO of past prices much like a fib "indicator" for example. No different in spirit than using Woodies CCI or a Williams% etc. to help determine where to buy. You are not using raw price data you are taking those prices and applying additional steps to them.
 
CoolHandLuke said:
To me you're lost in semantics. You draw an upward pointing strait line connecting the bottom of a few 5 min candles. That line forms arbitrary acute angle with the horizon the magnitude dependent of the steepness of of the upward trend. You then extrapolate that line to determine where to buy. The price your strait line instructs you to buy is simply a rough RATIO of past prices much like a fib "indicator" for example. No different in spirit than using Woodies CCI or a Williams% etc. to help determine where to buy. You are not using raw price data you are taking those prices and applying additional steps to them.
More...

Truely really doesn't matter what you use. Some years ago they had monkeys throwing darts at wall street journal page on stocks, and the monkeys did better than people in six months. When stocks going up, stocks rise. Using trendlines is easier as you can form rules of how much degrees of angle you won't use beyond, then you study for years on price patterns and swings of price to educate yourself if knowing what stage the trend is at right now, not all are the same. I am not huge into Elliot Wave, but I understand the concept of three wave in one direction, Waves 3 and 5 are best for confirmed trend, using trendlines is connecting starting point of wave 1 low and drawing through wave lows 2 and 4, taking price bounces past Wave 2 low for long positions. But trendlines are the easiest to understand, but much tougher to backtest unless you know how to do that type of coding.

Comparing "Woddie CCI"(What a joke) and Williams %R is tougher as you don't know into the future where it will be, then it is tougher to have well defined price patterns cause you don't know where the indicator will be one bar from now.

What most traders though don't have in their back pocket is "understanding" the whys. Why does price do what it does, most don't care and say price is 100% random. For one you have to look at Volume, volume of people using something, whether it is a newsletter(Robert Prechter's Elliott Wave Theorist), Fibs or trendlines, you get enough people who trade heaviest volume, they will cause the markets to go in same direction.
 
You must log in or register to reply here.
Back
Top