Hello, I am currently back-testing an algorithm on the ES and I would like to know how good/bad the liquidity is outside the regular hours when trading 20 contracts. Will LMT orders at the last traded price for example remain unfilled/partially filled more often than not ? Just to avoid the responses "newbie trading 20 contracts", I actually intend to trade 10 contracts at a time however the algorithm works by always being in a position so for example if I close a 10 contract long position, I will sell the 10 contracts that make up the long position and go short 10 contracts as well (so 20 contracts in total). During regular hours, I assume it's a breeze to get filled at the last executed price but I am very concerned about after/pre hours. The back tests shows good results on regular hours but the results are even better outside those hours. Thank you for any help is appreciated !!
most likely you need at least 500 contracts to move ES 0.5 point during the least liquid hours 6pm-1am I wouldnt worry about 20 contracts
This is a "feels" like answer. Since ETH SPY began being offered it's OK for a ten lot normally - the interaction between the two has helped "a bit".
Thank you for the reply ! In your experience, getting filled at the last executed price for 20 contracts shouldn't be a problem ? Looking the at the volume data, there are some times (around 1am-4am) where the minute volume is <20. I reckon my hypothetical 20 contracts limit order would be left unfilled or partially filled at best ?
Should be more than fine with liquidity, the question is the other assumption for the model, avg profit target? what happens if you factor in commissions/slippage of $10 or so etc.
slippage of 1 tick on each trade would destroy my model. That is why I am worried about getting filled at the last executed price. I base my back testing on that, the last traded price. If I always get filled at that price, I would be very confident in the model but I think the volume isn't there during the really dead periods (e.g 1-2 am)
If a tick kills it you might want to reconsider. I’m guessing your assuming “x” happens, place limit and you get filled. So now your at the back of the queue waiting for a fill, most likely it will trade through and against your position before you get filled. Never hurt I guess to try a couple single contracts. Generally, anyone I’ve seen ever try to model market making for ticks with just OHLC data usually finds out that small slippage and execution costs eat up any expected profit.
Well, maybe the reason you are seeing minute volume <20 contracts is because you are not there with your 20 contracts, providing the liquidity the market needs. Give it a whirl live, and see what happens!
go live with 1 or 2 micros and get a feel for what percent of the time you get filled and how the times you are not affect he model. if good move to 1 ES and progress. what software you use to design system?