Don't tell me then that TA doesn't work - page 19
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
You have to look at the cumulative BP of the OOS profit. The number of trades in the OOS does not matter. It is the number of OOS itself that matters.
Yes.
You have to understand that when you look at BP-profits, where the i-th element corresponds to the profit at time i. Then it does not matter for you how many deals were there.
And you can estimate total of such BPs by summing them up. Therefore for statistical significance it is necessary to have a decent number of such BPs.
You have to look at the cumulative BP of the OOS profit. The number of trades in the OOS does not matter. It is the number of OOS itself that matters.
On the contrary, it is the total number of deals that matters. You don't have to add up anything - just put it into the total state and count all the rest as usual (FF, MO, FS, etc.)
On the contrary, it is the total number of trades that counts
About the number of trades in this case is bullshit. And here's a simple example to prove it:
Obviously such a TS is working with a huge probability.
Now imagine that you do not know the length of OOS, but you would like to know on what length OOS your system is particularly good.
If you only look at OOS metrics such as PF, PV, etc, you will always be limiting yourself to a specific OOS length, since these metrics are dependent on the length of the OOS.
Or you can do much smarter. Take, for example, the interval of OOS month and add thousands of corresponding BP-profits of these intervals.
Then obtain the total BP-profits of the OOS. Then visualize this BP and see that during the first day, the total BP-profit goes up, and then similar to the SB.
Hence conclude that the steepest OOS length for you is one day. And note, no absolutely PF, FS and other crap depending on OOS length and number of trades.
By the way, I recall a very authoritative person saying that the only good filter is the one that improves performance, but does not reduce the number of deals (or insignificantly).
It's already allergic to authority references. Especially the anonymous ones. Can you make such a statement on your own behalf? Then we will discuss it.
Or bring your esteemed authority here. Otherwise, there is a lot of chatter in the bush around here...
You propose to count, for example, the total PF. Why engage in such a huge loss of information when you can sum the BPs and get everything in the palm of your hand?
It's all in the palm of your hand. How will you sum up - on one OOS 100 trades, on the other 10? Anyway, at adding a lot of important information will be lost, because there will only be an arithmetic mean at some moments. The spread relative to it will not be visible. This is important for analysis, though PF takes it into account in a simplified form. Sharps or Sortino are more accurate for this purpose, but PF will do.
About the number of trades in this case is bullshit. And here's a simple example to prove it:
That's funny :) Do that and your next trade will always be profitable, then optimize again ;)
It's already allergic to authority references. Especially the anonymous ones. Can you make such a statement on your own behalf? Then we'll discuss it.
Or bring in your esteemed authority. There's a lot of chatter in the bush around here...