Jump to content

Best time to short?


Recommended Posts

Roubini says "Don't blame me for today's rally. B)

 

THURSDAY, JULY 16, 2009

 

Roubini: "My Views Were Taken Out Of Context."

Posted by Tyler Durden at 5:26 PM

Does Crapvision now have to pay back all those who bought stock after being pounded by this misquote over and over as an epilepsy inducing Breaking News box? That, of course, assumes the powers that be will allow the market to move down even 1 point from the most recent artificially high level.

 

 

ZERO

Link to comment
Share on other sites

  • Replies 40
  • Created
  • Last Reply
Like TJ said a few weeks ago, the big bucks are in microsecond arbitrage algorithms, push-pulled with millions bought/sold daily. No net liquidity.

 

The "head and shoulders" pattern was megaphoned to anyone within earshot of Whiskey Haynes and Sooee Herrera. It was a brilliant suck in for all wannabe TA types who were blasted out of the RE flipper market. Those that had any money left, that is.

 

Now they don't.

Zactly.

 

Looks great on a chart.

 

So great that any moran will believe it -- without having any financial knowledge.

 

It even sells research reports.

 

Head and shoulders / rising wedge / bear flag whatever.

 

Not saying it's all crap, just that trading profitably is hard, and that trading profitably using with TA is still hard. :ph34r:

 

Edit: I like wedges, support and resistance, but haven't tested them.

Link to comment
Share on other sites

The question I was alluding to was whether TA worked or not, posed by rationalize in last night's M2M who suggested that it always resulted in losses over time as measured by testing various mechanistic systems. The arbitrage programs are a separate issue.

** Regular vanilla moving average trend following TA is not working in the backtests.

 

All I can say is that the inherent lag in MAs / EMAs means that you always miss the move.

 

And on a shorter timeframe, you catch the move, then whipsaw out too early.

 

So, no win either way.

 

The danger is that EMAs especially, or SMAs in spaghetti mode look freeking great on a trending chart.

 

The reality of trading with them mechanically -- entering every time the average say, and exiting every time they say -- is an average loss.

 

Now, if they are not being used mechanically, the WTF are they for?

 

How is the trading decision being made?

 

The line between analysis and gambling is crossed at that point IMO.

Link to comment
Share on other sites

I have had people tell me this works, though of course I don't believe it.

 

http://www.fapturbo.com/

Good example of backtesing misuse.

 

It's actually pretty easy to produce backtest results like those claimed.

 

Step 1: Get any data set.

 

Step 2: Code some plain valilla TA. Moving averages / RSI / Stochastics.. whatever. [can be done in excel +/-].

 

Step 3: Test across the whole data set, varying the TA input parameters, and the time period of the bars.

 

Result: Some funky combination will probably look fantastic.

 

Why is this? Well it's because a "curve" has been fitted to the data. Fitted to that specific sample of data.

 

What's the problem? Well, this says nothing about how the TA will perform on out of sample data.

 

i.e. How will it perform tomorrow? Probably not so well.

 

How to avoid this problem? Easy. Optomise the TA on one set of data, and then test it on another set of data...

 

Anything else is a scam for non-math types.

Link to comment
Share on other sites

** Regular vanilla moving average trend following TA is not working in the backtests.

 

All I can say is that the inherent lag in MAs / EMAs means that you always miss the move.

 

And on a shorter timeframe, you catch the move, then whipsaw out too early.

 

So, no win either way.

 

The danger is that EMAs especially, or SMAs in spaghetti mode look freeking great on a trending chart.

 

The reality of trading with them mechanically -- entering every time the average say, and exiting every time they say -- is an average loss.

 

Now, if they are not being used mechanically, the WTF are they for?

 

How is the trading decision being made?

 

The line between analysis and gambling is crossed at that point IMO.

 

Have you ever heard of a guy named K Wave Rider?

 

There are many variables and nuances in a chart. Chart reading is an art. It takes years of experience to learn how to do it well. Mechanistic systems based on automatic signals can't work because they are too rudimentary. I suggest you read Pring, or Edwards and Magee, and Hurst to get an idea of the basics of TA. You should also go back and read of of LeeWhee's posts and blog articles. Simply by using only long term support and resistance levels, he was able to trade successfully, rarely suffering a losing trade. K Wave is in a class by himself, but there's a lot you can learn from him, perhaps most importantly, studying charts incessantly for recognizable setups. That involves looking at moving averages, indicators, and price patterns on a variety of time frames on many charts to find one that has a high probability setup of a big move.

 

That's what TA is about. Could you write a program to do all that? I know that I couldn't but I suppose that there are geniuses out there who could. If they are successful at it, they sure aren't telling anybody. Why would they?

Link to comment
Share on other sites

There are zillions of these "systems" out there. Very easy to come up with. All you do is have your computer backtest a few thousand "systems" based on any random thing from buy on a certain hour or day and sell on another, to standard TA, to you name it. Then you find a few that "work" for a few weeks or months at a time or so. Just by random chance, a few of these "systems" will work profitably for a limited period of time; the probabilities of all 3 thousand systems or so losing is rather small. Maybe these systems only work for a few trades here and there and lose the rest of the time. Doesn't matter. The few times they work are the times you describe in your ads, and you pretend that the loser times and trades don't exist. But that's a lot of trouble, so I suspect most of the ad writers just outright lie.

Agreed.. Perfect example.

Link to comment
Share on other sites

...

That's what TA is about. Could you write a program to do all that? I know that I couldn't but I suppose that there are geniuses out there who could. If they are successful at it, they sure aren't telling anybody. Why would they?

They would tell everybody so that the system works.

 

The only reason that price moves, is beacuse everyone expects price to move.

 

If enough ppl "believe" in a system, the system works. [Think Black Scholes.]

 

The difficulty in writing a program to do this, is that the definition of resistance is whatever the gambler wants it ti be at the time.

 

"Oh, no, I meant resistance in another timeframe"....

 

"Oh, that trade doesn't count. It was a 'fake'"....

 

"Oh, false signal."....

 

"Oh, someone is 'manipulating'"....

 

"Oh, no, I'm trading on 'long term' resistance now"....

 

"Oh, no, I meant breaking through resistance, not bouncing off resistance"....

 

"Oh, no, well sometimes more retests of resistance means price will break through. Sometimes it's re-infocing the 'steel I beam'"....

 

Ever played blackjack?

 

Ever heard the BS that a problem gambler comes up with?

 

Edit: I agree that vanilla TA can look great on a chart. Trading with it in isolation would seem to be pretty risky.

Link to comment
Share on other sites

They would tell everybody so that the system works.

 

The only reason that price moves, is beacuse everyone expects price to move.

 

If enough ppl "believe" in a system, the system works. [Think Black Scholes.]

 

The difficulty in writing a program to do this, is that the definition of resistance is whatever the gambler wants it ti be at the time.

 

"Oh, no, I meant resistance in another timeframe"....

 

"Oh, that trade doesn't count. It was a 'fake'"....

 

"Oh, false signal."....

 

"Oh, someone is 'manipulating'"....

 

"Oh, no, I'm trading on 'long term' resistance now"....

 

"Oh, no, I meant breaking through resistance, not bouncing off resistance"....

 

"Oh, no, well sometimes more retests of resistance means price will break through. Sometimes it's re-infocing the 'steel I beam'"....

 

Ever played blackjack?

 

Ever heard the BS that a problem gambler comes up with?

 

"The only reason that price moves, is beacuse everyone expects price to move."

 

You're full of shit.

Link to comment
Share on other sites

I've been running this board for more than 8 years and have watched a number of people trade successfully using TA, in some cases for the entire time. It is not an easy discipline to master. Neither is playing the violin or composing music. But that does not mean it can't be done, and there are some who become virtuosos. The idea that you can build a simple rudimentary mechanized system that fails, and that that represents TA, is just idiocy. Rationalize is clueless about TA, so he's going to get a time out. Maybe he'll take the time to actually learn something, rather than just bloviating.

Link to comment
Share on other sites

had tp break out the ol' bizarro chart to check...if'n it don't hold here, 968 looks pretty doable

-------------------------

 

I LOVE it.... my wife always says I am going off the deep end when I occasionally stand in front of the mirror with my laptop upside down. I just tell her I am trying to get a different perspective... which is true :rolleyes:

Link to comment
Share on other sites

-------------------------

 

I LOVE it.... my wife always says I am going off the deep end when I occasionally stand in front of the mirror with my laptop upside down. I just tell her I am trying to get a different perspective... which is true :rolleyes:

 

 

She might think less crazy of you if you would put some clothes on first! :lol:

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Tell a friend

    Love Stool Pigeons Wire Message Board? Tell a friend!
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • ×
    • Create New...