For strategy 3, for the blended volatility estimate, for the first 10 years, do you use in sample fitted estimate for the entire dataset? Or median/mean of whatever data we have?
If you accuse me of in sample fitting on this thread again, you're out. Final warning https://github.com/robcarver17/pysystemtrade/blob/develop/sysquant/estimators/vol.py#L171 Mean so far Rob
Ouch. Not a great day of freedom for me. Portfolio is down about 80 BP. Nice Equity gains offset by FX and VIX losses. @globalarbtrader I have seen that you do not have any volatility positions on. Is this a coincidence or did you stop trading them? I restricted the volatility asset class to 4% weight but consider to move both VIX and V2TX to equities to dampen it even further. Volatility is definitely a winner but it's skew sucks.
Coincidence, I didn't stop trading them because I am not a pussy. VIX is flat, but VSTOXX has a short -10 forecast so if I had the capital I would have had a position on. 80bp seems okay vs 400bp in equities, no? My p&l is yo-yoing between flat and down -1.5% Rob
Yeah 400bp is quite disturbing. Interestingly, my V2TX forecast is in the same ballpark as yours but VIX hit it's record high of -10.3. The last 2 years it mostly has been -20 and I have always a position on. I do not trade the same strategies as you but have to look into it. Could be that I trade further out on the curve due to rolling criteria. My current holding contract is June 2025 which seems quite far out. Edit: Got it: The May contract was quite illiquid mid of february so my system rolled to the June contract. Seems like a sensible decision at the time.
I'm also down a bit, but not much, also had a V2TX short, -2 May contracts, but no stock positions of any kind at this moment (except for my long-term portfolio, which I don't want to look at, instead I'd rather look at the 'dividends' line of the fresh monthly statement and consider it a good news )
In my static system, (per Rob's work https://qoppac.blogspot.com/2016/03/diversification-and-small-account-size.html) I have a rough heuristic to include an instrument only if I can hold two contracts (after thresholding) at max signal strength. One issue that the most recent market turmoil revealed; was that being able to hold two contracts in table times, very quickly disappears in unstable times as vol spikes (capital falling didn't help). Was wondering about everyone thoughts if this issue was better handled 1) Ex-ante by having a bit stricter hurdle/heuristic for instrument inclusion. 2) Ex-post by removing instruments from system as vol spikes and their risk per contract becomes to great.
My .002c - I just use a diversified static set of assets that I think I can handle with my account and simply round to the nearest contract. I find this rounding approach tracks the original portfolio very closely. If a contract gets really big because of market moves e.g. cocoa/coffee you lose a bit of atomicity - but it's 'tolerable'.
Hello guys. I've been trading stocks following the most basic version of the Starter System: MAC16,64 as the opening rule and a trailing stop loss as the closing rule. Now I'm looking into improving the system by adding more rules and dropping the stop loss. I was planning to add a carry rule and allocate 50% weight to the MAC16,64 rule and 50% to the carry rule. But I read in chapter 11 in AFTS that the recommended allocation is 60% to trend following rules and 40% to carry rules. However, in chapter 16 in AFTS, trend is said to work poorly for equities and the best allocation is 10% to trend and 90% to carry. As I'm trading stocks only, should I adopt these weights? Which allocation are you using when trading equity futures? This leads to another question: If I were to allocate 90% to carry, would I be able to drop the stop loss from the system? LT suggests that the stop loss should be kept if more than 50% of your opening rule weights are in carry. Thanks.