Predicting the characteristics of weekly bars (championship). - page 2

New comment
 
sibirqk:



Well, if you don't try it, it definitely won't work. If there is a signal without noise on forex, then it should appear best on a weekly timeframe - it's not five-minute or M30, where the next bar is absolutely unpredictable. On the weekly timeframe the difference between low and high is more or less the same. Again, the trends are relatively smooth - using intuition, supplementing it with ishimics, fleas, mcd's and other stuff, one can learn to accurately guess the weekly range. I think so.
Just the shorter the prediction period - the more accurate the prediction.
 
sibirqk:

I have invented a useful and entertaining competition theme - every week, anyone can post a forecast of the parameters of a future weekly candle -high, low, open, close. Monday is a deadline. The number of pairs, the number of parameters in a pair, in the sense that it's not necessary to forecast all four parameters, it is possible to forecast two or even one. After the close of the next week the forecast is evaluated in points. Scores of each participant are accumulated, and after some period it will be possible to objectively estimate who and how accurately the future can be seen. The same genius traders-hedgers, of which there are a dense number in this forum, can create a robot that will automatically predict the characteristics of a weekly candle, and the author will post them in a branch. If the forecasts are successful, it will advertise their creation. In general, something like a championship ersatz, with a prize in the form of the public worship in narrow circles.

It is clear, that the system of scoring must be transparent, understandable, universal (the same for all pairs) and independent of the number of predictions - i.e. if few predictions are made, but more exactly, then the total score is more than if many inaccuracies.

As a starting point, I propose the following evaluation method:

The most accurate and the most inaccurate predictions are evaluated with the same number of positive and negative points. This is to sift out random good and bad predictions. Minimal/minimal points are set to 10, with an appropriate sign.

There is a close price of week and the difference between it and future actual value is 100%. If it's within +/- 10 percent, you get the maximum positive point, if it's below 100 percent, you get the maximum negative point and, somewhere in the middle, you get zero.

For example, the euro closing price was 1.0912, last week's high was 1.0975, the difference is 63 points. If someone forecastedhigh- 1.0970 or 1.0980, they would get 10 points, and if they forecast 1.1040 and beyond - 10 points.

It is clear that the scoring system is approximate so far, and can be improved from here.

A Siberian and a chimpanzee do not fit together. If we had a Kenyan or an Angolan + chimpanzee it would be better. But a Siberian is better suited for a bear.)))

 
for the week of Feb 8: USDJPY - inverted hammer with a wick around 118, over 119 unlikely. In short - I expect close at the open, short shadow and long wick.
 
Alexey Volchanskiy:
Just the shorter the prediction period, the more accurate the prediction.

If you count in percentages - actual bias/forecast, I think this is unlikely to be the case.

Maybe you can demonstrate it? For example on H4, on some currency pair?

 
khorosh:

A Siberian and a chimpanzee do not go together. If it were a Kenyan or Angolan + chimpanzee, it would be different. But a Siberian is more suited to a bear.)))

A bear is so clichéd. You should get rid of stereotypes, rise above the clichés. Besides, this year is a monkey year. In general - a monkey as a premonition, or as a state )))). And the bear, in these times, is almost a politician))))
 
Maxim Kuznetsov:
for the week of Feb 8: USDJPY - inverted hammer with a wick around 118, over 119 unlikely. In short - I expect close at the open, a short shadow and a long wick.
I am thinking somehow similar. What would you say about the figures? About the shadow and the wick?
 

AUD/USD: Hight 0.77545 Low 0.76137

BRN: Hight 71.30 Low 68.36

Current closing price: AUD/USD:0.76960, BRN:69.80

 
Two years have passed and the predictions are still not coming true
 
Morexod:
Two years have passed and the forecasts have not come true

The forecasts here were for a week, not two years.

 
Aliaksandr Maksimau:

AUD/USD: Hight 0.77545 Low 0.76137

BRN: Hight 71.30 Low 68.36

Current closing price: AUD/USD:0.76960, BRN:69.80

Oh, this thread has appeared and I had already forgotten about it.

Almost agree on the Aussie - maybe would have shifted down a bit - 20 pips, Hg and Lw. The week ended with the American stock slump, I am afraid that this week the drop will continue, consequently the commodities will start to fall and as a consequence the Aussie will suffer.

As for oil, it may take flight this week as USA threatens to impose sanctions on Iran, plus Venezuela is about to collapse, which may well push oil over 75. In short, I would put Brent at Hg -75 Lw - 68 for the week.

12345678
New comment
Reason: