The Skill Check | Page 5 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

No skill in trading! 107 replies

Poll: Is successful trading more dependent on skill than luck? 149 replies

The most important technical skill to learn 12 replies

Need luck or skill for trading? 232 replies

Need suggestion to improve skill 35 replies

Attachments: The Skill Check
Exit Attachments

The Skill Check

  • #81
  • Mar 16, 2018 8:26pm Mar 16, 2018 8:26pm
Quoting !iii!
The "problem" is the proximity of the daily open prices to the pivot level, which doesn't give much "room" to work with: 5% <= 1.4 pips 10% <= 2.7 pips 15% <= 4.1 pips 20% <= 5.7 pips 25% <= 7.0 pips ... 50% <= 14.7 pips
You could calculate the pivot for each new M1 candle. For instance, at midnight you calculate the pivot based on the OHLC of the recently closed day. At 00:01 you use the close of the M1 candle as well as the highest high and lowest low of the range from 00:01 of previous day to 00:01 of the current day, etc.

This way you may have much more trading opportunities and thus you could just wait for the pivot points with a relatively high distance. However, the problem here is that many consecutive pivot points may not be hit.
  • #82
  • Mar 16, 2018 9:04pm Mar 16, 2018 9:04pm
Quoting AntiCre
{quote} You could calculate the pivot for each new M1 candle. For instance, at midnight you calculate the pivot based on the OHLC of the recently closed day. At 00:01 you use the close of the M1 candle as well as the highest high and lowest low of the range from 00:01 of previous day to 00:01 of the current day, etc. This way you may have much more trading opportunities and thus you could just wait for the pivot points with a relatively high distance. However, the problem here is that many consecutive pivot points may not be hit.
Now we're talking
  • #83
  • Edited 1:01pm Mar 17, 2018 10:27am | Edited 1:01pm
Quoting AntiCre
{quote} You could calculate the pivot for each new M1 candle. For instance, at midnight you calculate the pivot based on the OHLC of the recently closed day. At 00:01 you use the close of the M1 candle as well as the highest high and lowest low of the range from 00:01 of previous day to 00:01 of the current day, etc. This way you may have much more trading opportunities and thus you could just wait for the pivot points with a relatively high distance. However, the problem here is that many consecutive pivot points may not be hit.
This is how the proximity of open prices to the continuous daily pivot looks, when using a H1 resolution - just like any other oscillator. As expected, the hit rate remains high, but so do the consecutive misses (I saw streaks of 24 misses on 1,000 trades). Even if we enter only upon certain thresholds, I can't see, how the pivot stats with the high hit rates help us now.
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1.png Size: 57 KB
Augmenting Intelligence
  • #84
  • Mar 18, 2018 12:15pm Mar 18, 2018 12:15pm
Sticking with the calculation of continuous pivot levels, I simulated

opening trades
a) from the open price towards Pivot, no hard SL (i.e. every bar), and
b) from sigma 1 and 2 excursion levels, but with additional filters being

  1. Reward-to-risk at 2 (i.e. Distance to Pivot = 2 x SL)
  2. Open price to SL >= 5 pips, and
  3. The period high or low having happened before the actual open price -> i.e. the open price has a floor / ceiling

and closing trades at the Pivot (=profit), SL or with expiry after 1 day (=loss).

The hit rates dropped significantly from a) to b), and, unless I have a bug somewhere, the profit:loss stats didn't suggest a statistical edge, either.

I used 10,000 bars and with every sigma trades dropped tenfold and hit rates halved.

Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1.png Size: 58 KB
Augmenting Intelligence
  • #85
  • Mar 18, 2018 1:59pm Mar 18, 2018 1:59pm
Quoting !iii!
Sticking with the calculation of continuous pivot levels, I simulated opening trades a) from the open price towards Pivot, no hard SL (i.e. every bar), and b) from sigma 1 and 2 excursion levels, but with additional filters being Reward-to-risk at 2 (i.e. Distance to Pivot = 2 x SL) Open price to SL >= 5 pips, and The period high or low having happened before the actual open price -> i.e. the open price has a floor / ceiling and closing trades at the Pivot (=profit), SL or with expiry after 1 day (=loss). The hit rates dropped significantly from...
I really like how you are visualizing the distance with the rolling pivots. Creates a very interesting histogram.

I think it does so poorly because the stops are super close. I'd try using a stop that is the same length as your target or 2 times greater than the distance to the target.

Also, I'd be interested to see the basic pivot hit rates from each sigma level using no stops.
  • #86
  • Mar 18, 2018 3:37pm Mar 18, 2018 3:37pm
No stops
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_nostops.png Size: 76 KB
Augmenting Intelligence
  • #87
  • Mar 18, 2018 3:37pm Mar 18, 2018 3:37pm
1R
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_1R.png Size: 76 KB
Augmenting Intelligence
  • #88
  • Mar 18, 2018 3:37pm Mar 18, 2018 3:37pm
0.5R
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_05R.png Size: 77 KB
Augmenting Intelligence
  • #89
  • Mar 18, 2018 3:38pm Mar 18, 2018 3:38pm
All of the above under the assumption of no bugs, of course..
Augmenting Intelligence
  • #90
  • Mar 18, 2018 3:48pm Mar 18, 2018 3:48pm
A close up for visualization of histogram and on chart rolling pivot
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_closeup.png Size: 58 KB
Augmenting Intelligence
  • #91
  • Mar 18, 2018 3:50pm Mar 18, 2018 3:50pm
Nice progress! How have you calculated the sigma, I mean over which period?
  • #92
  • Mar 18, 2018 4:09pm Mar 18, 2018 4:09pm
Quoting AntiCre
Nice progress! How have you calculated the sigma, I mean over which period?
Over all the 10,000 samples.
Augmenting Intelligence
  • #93
  • Mar 18, 2018 4:10pm Mar 18, 2018 4:10pm
From 9.8.2016 up to today.
Augmenting Intelligence
  • #94
  • Mar 18, 2018 4:24pm Mar 18, 2018 4:24pm
115 samples to validate the figures - they appear to be correct
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_115samples.png Size: 52 KB
Augmenting Intelligence
  • #95
  • Mar 18, 2018 4:38pm Mar 18, 2018 4:38pm
Quoting AntiCre
Nice progress! How have you calculated the sigma, I mean over which period?
Since the true range differs significantly between the hours of the day, the calculation of sigma could be based on just all values of the same hour. However, I don't think that it will change the overall picture ...
  • #96
  • Mar 18, 2018 4:45pm Mar 18, 2018 4:45pm
Quoting !iii!
115 samples to validate the figures - they appear to be correct {image}
Could you please explain what this shows exactly? Looks promising
instagram: stepbystepforex
  • #97
  • Mar 19, 2018 4:30am Mar 19, 2018 4:30am
Quoting AntiCre
{quote} Since the true range differs significantly between the hours of the day, the calculation of sigma could be based on just all values of the same hour. However, I don't think that it will change the overall picture ...
It didn't...
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDH1_periodicsigma.png Size: 71 KB
Augmenting Intelligence
  • #98
  • Mar 19, 2018 4:39am Mar 19, 2018 4:39am
...however, going with periodic sigmas only halved (instead of dropping tenfold) the opportunities per excursion level while the hit rates remained above 50%.
Augmenting Intelligence
  • #99
  • Mar 19, 2018 4:47am Mar 19, 2018 4:47am
Quoting rafael5643
{quote} Could you please explain what this shows exactly? Looks promising
For now it shows, that Pivot levels remain a Fata Morgana.
Augmenting Intelligence
  • #100
  • Mar 19, 2018 10:29am Mar 19, 2018 10:29am
Given the proximity to the pivot level for 50% of all hits being the main challenge to recover floating losses from previously missed pivot levels, one could try an Anti-Martingale approach: instead of focusing on trending streaks for double downs - i.e. the period of consecutively missed pivots -, we do a double up when a pivot is hit or when the hit was particularly small. Or we do it randomly.

Any ideas are welcome.
Augmenting Intelligence
0 traders viewing now