You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I am toying with the idea how to size up or size down an alpha based on its performance or Bayesian Consistency or some kind of performance measure.
My thought is that we could have a standard size risk reversal with a 30 delta target and and then increase the delta of the risk reversal if the alpha fell below a consistency level.
How could we accomplish this?
Dmitry, you were working on a ML sizing alpha, would that concept be applicable.
Alex, could you think of a V2 implementation that would be conscious of how the particular alpha was performing and then adjusted the deployed risk reversal to an appropriate delta?
thx
The text was updated successfully, but these errors were encountered:
I have done some test of concept work on the idea above an one implementation which may work is to build a SmartEXO in V2 that is composed of 2 risk reversal. Let consider the bull hedge case. During bullish phases in the underlying futures we would run a 20% hedge risk reversal and during bearish phases we would run a 80% delta risk reversal. I used a simple 10day Keltner channel on the 6B for this example. On days when the center channel was rising we hold a 20% delta risk reversal and on days when then center channel is not rising we hold the 80 delta risk reversal. When the condition changed from rising to falling center Keltner channel i give the time series the previous days weight so there is not a look ahead bias.
Here are the 6B futures alphas i used in the example.
Sent from my iPhone
On Oct 2, 2017, at 3:00 AM, alexveden <notifications@github.com<mailto:notifications@github.com>> wrote:
can we build a index hedge in V2 that works like this proposal?
Yep, absolutely possible.
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub<#57 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ARobLOksqJfIrZyXuqBfnoBkenVwlBxMks5soLRCgaJpZM4PntHw>.
I am toying with the idea how to size up or size down an alpha based on its performance or Bayesian Consistency or some kind of performance measure.
My thought is that we could have a standard size risk reversal with a 30 delta target and and then increase the delta of the risk reversal if the alpha fell below a consistency level.
How could we accomplish this?
Dmitry, you were working on a ML sizing alpha, would that concept be applicable.
Alex, could you think of a V2 implementation that would be conscious of how the particular alpha was performing and then adjusted the deployed risk reversal to an appropriate delta?
thx
The text was updated successfully, but these errors were encountered: