 |
Xspurt
Registered: Jun 2009
Posts: 1781 |
08-20-12 03:23 PM
Quote from TheBlackHand:
So let me get this straight:
1. TA cant be tested.
2. TA cant be evaluated.
3. TA has infinite different inputs.
4. No one person can test all of TA.
I love that first one sooooooo much!!
TA CAN NOT BE TESTED
So, how the hell can you be sure something works, IF, as you ADMIT, it cant be tested!!
And I thought back testing was the cornerstone of what yoy TA junkies believed in.
Thank you so much for being man enough to admit what the others cant see.
At last there is progress after weeks and weeks of debate in these various threads.
YOU SIR DESERVE A MEDAL!! A big shiny gold one too! 
Like I say, two people look at a chart and one sees the details and one can't connect the PA with the lines.
It's the same as in reading. Here is an example of not understanding what the word ALL means and hence the conclusions are, well, funny.
The reading medal goes to you TheBlackHand
I can only speak for myself on the next matter, but as far a back testing is concerned, I have never back tested anything. I have forward test for almost 30 years as I am not a great believer in back testing.
Really TheBlackhand, you have a weird concept of TA so it is understandable why TA seems weird to you.
|
| |
|
Edit/Delete • Quote • Complain |
logic_man
Registered: Oct 2010
Posts: 1489 |
08-20-12 03:45 PM
Quote from Xspurt:
Like I say, two people look at a chart and one sees the details and one can't connect the PA with the lines.
It's the same as in reading. Here is an example of not understanding what the word ALL means and hence the conclusions are, well, funny.
The reading medal goes to you TheBlackHand
I can only speak for myself on the next matter, but as far a back testing is concerned, I have never back tested anything. I have forward test for almost 30 years as I am not a great believer in back testing.
Really TheBlackhand, you have a weird concept of TA so it is understandable why TA seems weird to you.
While it is true that ALL possible TA combinations cannot be tested, the SPECIFIC ones any given trader tries to use can be tested.
If they don't work, move on to the next set of combinations.
|
| |
|
Edit/Delete • Quote • Complain |
TheBlackHand
Registered: Jun 2012
Posts: 221 |
08-20-12 03:46 PM
Quote from Xspurt:
It's the same as in reading. Here is an example of not understanding what the word ALL means and hence the conclusions are, well, funny.
The reading medal goes to you TheBlackHand
Indeed.
He said: 'of course ALL ta can not be tested'
Implying perhaps some part of it can be tested, and part therefore cant; Which part? A big part?, a little part?, the most part? It would seem as if the part is as mysterious in definition as TA itself!
One thing that has not been given in this tirade of posts is any objective testing on TA.
Will anyone step up the challenge and point to some HARD EVIDENCE THAT TA WORKS THROUGH TESTING OF A FEW CONCEPTS?
(enjoy the ensuing silence....)
|
| |
|
Edit/Delete • Quote • Complain |
TheBlackHand
Registered: Jun 2012
Posts: 221 |
08-20-12 03:49 PM
Quote from logic_man:
While it is true that ALL possible TA combinations cannot be tested, the SPECIFIC ones any given trader tries to use can be tested.
If they don't work, move on to the next set of combinations.
Which set?
The infinite set?
And if you do find one that works (probably a random sequence of heads being flipped in the infinite expanse of TA outcomes), what happens when the market cycle changes?
|
| |
|
Edit/Delete • Quote • Complain |
kut2k2
Registered: May 2005
Posts: 3144 |
08-20-12 03:51 PM
Quote from TheBlackHand:
So let me get this straight:
1. TA cant be tested.
Wrong again, or at least misleading. In order for a particular type of TA to be tested, it must be known.
So:
Known TA can be tested. Unknown TA (i.e., the vast majority of TA) obviously cannot be tested.
You keep wanting to generalize from a tiny subset of TA (aka TA in the public domain) to all possible TA. Sad.
|
| |
|
Edit/Delete • Quote • Complain |
logic_man
Registered: Oct 2010
Posts: 1489 |
08-20-12 03:56 PM
Quote from kut2k2:
Wrong again, or at least misleading. In order for a particular type of TA to be tested, it must be known.
So:
Known TA can be tested. Unknown TA (i.e., the vast majority of TA) obviously cannot be tested.
You keep wanting to generalize from a tiny subset of TA (aka TA in the public domain) to all possible TA. Sad.
It is ironic the number of holes in these people's arguments. It's almost as if they've never learned how to do "A", much less "TA".
I don't even use TA (or, I use a subset that is algorithmic, so no charts or indicators for me), but I don't go around spouting that it's impossible on the basis that I can't make it work.
|
| |
|
Edit/Delete • Quote • Complain |
| Receive
an email whenever a new post is added to this thread by subscribing
to it. |
|
|
|
|
