Я сегодня поиграл с данными и остался неудовлетворен. Сигнал есть, но шума больше. Четко установить связь между формой кривой и аномалиями роста XIV пока не могу. Но еще попробую.
Possibly an exercise in self-deception but the steps are as follows
1. For each day, compute dF=F2-F1 (two nearest VIX futures), dVIX (change in VIX) and change in the value of XIV since yesterday, dXIV.
2. Run a linear regression dXIV against dVIX over last 2 months.
3. Compute excess return of dXIV every day vs. exact linear dependence on VIX.
4. Compute moving averages of those excess returns and dF
5. Compute correlations of those MAs
Result: the correlation had a few dips (once into deep negative) but mostly stays to the North of 70%. Didn’t really think much. Just wanted to back up my idea that positive dF gives a boost to XIV.
no subject
Date: 2015-09-19 03:59 am (UTC)no subject
Date: 2015-09-19 05:21 am (UTC)а на чем вы с данными играете?
no subject
Date: 2015-09-19 02:07 pm (UTC)no subject
Date: 2015-09-23 08:21 pm (UTC)1. For each day, compute dF=F2-F1 (two nearest VIX futures), dVIX (change in VIX) and change in the value of XIV since yesterday, dXIV.
2. Run a linear regression dXIV against dVIX over last 2 months.
3. Compute excess return of dXIV every day vs. exact linear dependence on VIX.
4. Compute moving averages of those excess returns and dF
5. Compute correlations of those MAs
Result: the correlation had a few dips (once into deep negative) but mostly stays to the North of 70%. Didn’t really think much. Just wanted to back up my idea that positive dF gives a boost to XIV.
no subject
Date: 2015-10-20 06:29 pm (UTC)посмотрим...
no subject
Date: 2015-10-22 12:14 pm (UTC)no subject
Date: 2015-10-22 04:11 pm (UTC)