The Really Useless Thread | Page 617 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

The Really Useless System 25 replies

Bloomberg: VIX Indicator useless for forecasting the direction of equity prices 0 replies

The Really Really Boring Forex System 97 replies

Attachments: The Really Useless Thread
Exit Attachments

The Really Useless Thread

  • #12,321
  • Sep 13, 2011 3:30pm Sep 13, 2011 3:30pm
this is my e/u.. second try to add one. so far it works much better as a scalp. I mentioned earlier 3730 and if that goes then 38xx is in sight (3830 ?) if thats the case it will get stopped out. SL is tight..

Attached Image (click to enlarge)
Click to Enlarge Name: eu.gif Size: 23 KB
  • #12,322
  • Sep 13, 2011 3:30pm Sep 13, 2011 3:30pm
Quoting OldQuant
Again, just a suggestion.
And a damn fine one!

Bookmark ladies and gentleman!
  • #12,323
  • Sep 13, 2011 3:30pm Sep 13, 2011 3:30pm
Quoting darimik
whats the meaning of 2 ln 2 -1 ? since its a constant, why not keep it that way...
Yes, it is a constant; in the class that this code came from, it gets set once in the class' constructor. The origin of it is in the maths behind the "range" that come from non-parametric statistics. You could read the original paper, but your point is right on: it's a constant.
  • #12,324
  • Sep 13, 2011 3:32pm Sep 13, 2011 3:32pm
Quoting cloggie
And while you are at it google DU volatility.....
even though not directed at me, but noted... thirsty for knowledge..
  • #12,325
  • Sep 13, 2011 3:32pm Sep 13, 2011 3:32pm
Quoting OldQuant
Here's a FWIW suggestion. An instantaneous estimator of the volatility in a bar (of any TF) is:

.GKvol = 10000 * Sqr(0.5 * (.hi - .lo) * (.hi - .lo)
- twoLn2_m_1 * (.op - .cl) * (.op - .cl))

Here twoLn2_m_1 = (2*Ln(2) - 1), hi, lo, op, cl are self explanatory. The 10000 scales it up to pips, and Sqr is square root. This is called a Garman-Klass estimator, which you can Google.

Anyway, keep a moving average of those. Then you can set a stop according to some multiple of that value (fairly big, like 6 or so)...
thanks OQ, yes an arbitrary "one size fits all" doesn't really work..

slightly concerned that i do actually understand that equation...

or maybe just use something like a multiple of atr?
  • #12,326
  • Sep 13, 2011 3:33pm Sep 13, 2011 3:33pm
Quoting OldQuant
Yes, it is a constant; in the class that this code came from, it gets set once in the class' constructor. The origin of it is in the maths behind the "range" that come from non-parametric statistics. You could read the original paper, but your point is right on: it's a constant.
thought it may have been something relating to the timeframe.had a few beers , so hard to put it around my head atm.
  • #12,327
  • Sep 13, 2011 3:35pm Sep 13, 2011 3:35pm
Quoting FibbinArchie
thanks OQ, yes an arbitrary "one size fits all" doesn't really work..

slightly concerned that i do actually understand that equation...
It's pretty simple, really. What's interesting about it is what happens when op=cl, or a total pinbar. Then the second term is zero, and the measure has its max value; "pure volatility." When op=hi and lo=cl, or vice versa, then the measure has it's min value: a bar with no wicks is closest to pure signal.
  • #12,328
  • Sep 13, 2011 3:37pm Sep 13, 2011 3:37pm
Quoting hkkhanus
might be party time for some
Indeed...
Attached Image


Attached Image
  • #12,329
  • Sep 13, 2011 3:39pm Sep 13, 2011 3:39pm
Quoting OldQuant
It's pretty simple, really. What's interesting about it is what happens when op=cl, or a total pinbar. Then the second term is zero, and the measure has its max value; "pure volatility.

Correct me if I am wrong, but if my memory serves me right don't you use 70% if the bar has no body? And 33 with no wicks?
  • #12,330
  • Sep 13, 2011 3:43pm Sep 13, 2011 3:43pm
Quoting OldQuant
It's pretty simple, really. What's interesting about it is what happens when op=cl, or a total pinbar. Then the second term is zero, and the measure has its max value; "pure volatility." When op=hi and lo=cl, or vice versa, then the measure has it's min value: a bar with no wicks is closest to pure signal.
is it not possible that the hi and lo are concealing what has actually taken place during the bar period?
  • #12,331
  • Sep 13, 2011 3:46pm Sep 13, 2011 3:46pm
Quoting cloggie
Correct me if I am wrong, but if my memory serves me right don't you use 70% if the bar has no body? And 33 with no wicks?
Very close to that, yes! In the first case, it would be sqrt(.5) = .707, and in the second sqrt(.5 - (2 ln(2)-1)) = .337. (just did em on a calculator).
  • #12,332
  • Sep 13, 2011 3:51pm Sep 13, 2011 3:51pm
Quoting FibbinArchie
is it not possible that the hi and lo are concealing what has actually taken place during the bar period?
Sure, but it that's a concern, lower tf measures can be used. These things are "averagable", so the average of the values of 5 1M should, over time, be close to the values from one 5M that wraps those five minutes' values.

EDIT: The average of those 5 1M measures would have to be scaled up by a factor of sqrt(5) to compare with the 5M value, though. Vol measures are always related by square root of time. And that, BTW, is where the whole notion of "noise trading" comes from; as you drop tf, the vol doesn't drop as fast as the time frame, and the signal-noise ratio goes to hell.
  • #12,333
  • Sep 13, 2011 3:51pm Sep 13, 2011 3:51pm
Quoting kenfen69
Amen bastard!



Without trying to sound like a big-old cheese-ball...I consider it an honor, a privilege and a pleasure to be able to shoot the breeze and talk a little shop with everyone here on a daily basis. I won't be giving up on the this thread anytime soon, mate...it is a bit of a special place to me. I enjoy my time here...and hope others do, too.

Its like a virtual pit...a virtual bar...and occasionally, a virtual "horse" all rolled into one! What more could you want? A cup of coffee like PTL in the morning? Well...we've...
why don't you buy him a dinner and get a room.
those who can, do. those who cant, talk about those who can
  • #12,334
  • Sep 13, 2011 3:54pm Sep 13, 2011 3:54pm
Quoting OldQuant
Sure, but it that's a concern, lower tf measures can be used. These things are "averagable", so the average of the values of 5 1M should, over time, be close to the values from one 5M that wraps those five minutes' values.
cool, thanks OQ..
  • #12,335
  • Sep 13, 2011 3:57pm Sep 13, 2011 3:57pm
Quoting gatorinla
why don't you buy him a dinner and get a room.
Hahahahahaha...you bastard!

Who...PTL? Do you REALLY know where is right now?

  • #12,336
  • Sep 13, 2011 3:59pm Sep 13, 2011 3:59pm
Quoting gatorinla
why don't you buy him a dinner and get a room.
I miss PTL because Ken doesn't post those babes for any of the rest here.
  • #12,337
  • Sep 13, 2011 4:02pm Sep 13, 2011 4:02pm
gee we closed the est one at 3678.... decision time. h closes above bullish. below bearish... below is 3366 is another above 3686 is another..
those who can, do. those who cant, talk about those who can
  • #12,338
  • Sep 13, 2011 4:06pm Sep 13, 2011 4:06pm
The Dutch Finance Minister said they should prepare for orderly default, whatever that entails.

Which reminds me, I have to go get tobacco. It's becoming harder and harder to get good Dutch half-zware tobacco. Now the law is it can't be sold over internet; the merchant has to be able to see the customer. Personally, I think that's illegal. What if the merchant is blind?

As soon as this all restarts in 9 minutes, I'm out.
  • #12,339
  • Sep 13, 2011 4:17pm Sep 13, 2011 4:17pm
Quoting gatorinla
oil poked above a dt (wick).. eu back to 3800/3841.. is definitely not out of the question. will the mid hold at 3733.5..
cnbc had the oil experts on today... they said oil had crossed 90 and it meant bullish... i don't have a seat on the commodity board.. so my # being so close may be off.. anyway... i dispute their claim.... and with my bearish outlook on the gbp. of course i could be wrong.. but i will treat oil as a tt... and bearish.. on it now and still gbp .. i have a bullish formation on the euro targeting 3800/41 but w/o oil, it will be less bullish... my oil close is 89.94.. just below 89.95. 89.96 would have persuaded me for bulls.. it was just that close. this makes me less bullish on the euro then i was before this.
those who can, do. those who cant, talk about those who can
  • #12,340
  • Sep 13, 2011 4:20pm Sep 13, 2011 4:20pm
Quoting OldQuant
The Dutch Finance Minister said they should prepare for orderly default, whatever that entails.
But speaking of Greece, correct? I think the headline was misinterpreted earlier in the day...and it quickly had to be clarified.

Quoting OldQuant
Which reminds me, I have to go get tobacco. It's becoming harder and harder to get good Dutch half-zware tobacco. Now the law is it can't be sold over internet; the merchant has to be able to see the customer. Personally, I think that's illegal. What if the merchant is blind?
What kind...Druma?
210 traders viewing now