sorry... do you mean vega as opposed to gamma? Market has only gone one way.
Fully automated futures trading
- Thread starter globalarbtrader
- Start date
Oh, looks like something happened in the world while I slept vix is +40% 
I'm strangely up by an insignificant amount.. the system bought back 3 v2tx and sold one mxp overnight, but I still lost on them, but for some reason I'm short 2 cattle contracts which made money compensating for the loss
I'm strangely up by an insignificant amount.. the system bought back 3 v2tx and sold one mxp overnight, but I still lost on them, but for some reason I'm short 2 cattle contracts which made money compensating for the loss The limit orders - yes, I do it to try to capture bid-ask spread (it's a simple algo, thanks to the Rob's post
). Didn't analyze the results in prod yet, in paper it saved some money but I guess ib's paper execution isn't real. The other part - the whole event-driven thing is just because I felt it was a more natural approach from software perspective.. But its definitely more complicated and probably unnecesary. One posdible benifit is that the system reacts quicker to spikes like today. Not sure if it makes a diffetence long-term though..The limit orders - yes, I do it to try to capture bid-ask spread (it's a simple algo, thanks to the Rob's post
IB's paper execution is reasonably good and often pessimistic. Not saying you can rely on it of course, like combos are completely unrealistic fills but vanilla stuff is decent.
How do we improve execution for combos without hft infra?
Initially the week started about +2% due to the strong rise in GC, compared to the end of last week. Bonds (ZT, ZN, 3KTB, also GE) are up, but equity down (EU600, MES). Meanwhile are things going down a bit. Now I'm at about -2% compared to the end of last week. The volatility has clearly gone up since the weekend and several of my open positions had their position size reduced.
I am not sure what IB paper execution does, but limit orders are very hard to simulate if you take into account the position in queue for price/time priority order books. If you assume that you are first in the queue, you always gonna make money when showing passive orders. If you assume that you are the last in the queue, you never gonna make any money (the only time you get filled is when the whole level is taken out).
My prior is that most attempts at providing liquidity without HFT infrastructure are a losing proposition - i.e. you will get picked off by the faster guys and it will cost you more than aggressive execution.