Very helpful. You've explained what I've been observing (I'm kind of new at this so not really understanding what I was seeing until you explain it). Can you explain the advantages and disadvantages of putting in a programmed order to buy/sell at 9:30:01 @ market instead of MOO? Which would give a better price and under what circumstances? Thanks
I would strongly advise against placing a market order at 09:30:01. In the 1st few minutes of trading the spreads can be very wide on all but the most liquid stocks. I would either avoid the open altogether or submit orders for the opening cross. The opening cross typically represents about 1 to 2% of the total daily volume in a stock. I would limit your order to a small fraction of this otherwise you will pay slippage by pushing the opening price up or down with your buy/sell order. You could also bid/offer with a LMT order at 09:30:01 but the problem with this is that even though you may "capture" the spread, the spread at the open is wide for a reason... there is a lot of informed trading going on and so you will find that when your LMT order to buy/sell is filled, the price will often continue downward/upwards.
Thanks for this explanation, your logic is very sound. I trade an EOD system and often I'm not at my desk at the open (I'm often at work), so I can't watch the open and try to time a trade. So I've been using market open. From what you've said it would be better for me to use market on open rather than a market order at 9:30:01. What about MOC? Would that be the same? I've been using those as well. I'll glance at my trades towards the end of the day and if I need to exit I'll set a MOC order. The only problem is sometimes a stock will reverse in the last half hour (I think the cutoff for MOC is 20 minutes but I'm not sure) and I can't cancel my MOC order. Last week I had to put in another buy order to buy back what the MOC was going to sell. That's frustrating. SO I was thinking better to program a sell order for 1 minute before close that way I can cancel it if I need to. But I think it might give a worse price.
Yes, definately best to use MOO. Keep your order to less than 5% of the average opening cross volume and you shouldnât influence the price too much. As for the close. You could use a LOC order, so if the price falls in the last 20 mins it wonât get filled. If your âglancing at your trades towards the end of the dayâ then iâll bet it makes no difference if you execute 1 hour before the close, half hour before the close or at the actual official close price. I once wanted to trade at the exact close price (as I had tested systems using daily OHLCV data), but analysis using intraday data showed almost no difference in performance. MOC has the same advantage to MOO in that you get a single cross price and so donât pay the spread. But obviously if you trade too big a volume youâll influence the price giving slippage. Because MOC imbalances get published heading into the close you can generally trade far more volume at the close without moving the price because other traders step in and provide liquidity to the published imbalance.
I just recently got set up with intraday data so I'm going to do my backtesting and check to see if an exit at close, 30 minutes before, 1 hour, etc. which is best. Up until now I was going for close because that's the data I had for backtesting. How do I get the opening cross volume? I've been trying to determine how much volume I can have without impacting price. So far I've settled on 0.25% of daily volume. But in several cases I have detected differences in my fills and the trading done at 9:30.
As far as I know you cannot get advance figures for the current open as you can continue to submit orders until 09:28:00 for nasdaq and right up to the open for NYSE - perhaps Don can confirm this. However, you can download historical data from http://emi.nasdaq.com/?tab=openclose and by downloading tick/minute data for the NYSE from sites like opentick (use the open price of the 1st 1 minute bar and you can assume that 90% of the volume going through in this minute occurred at the cross), likewise for nasdaq as cross occurs randomly between 09:30:00 and 09:30:30. In my experience the opening cross volume represents about 1 to 2% of average daily volume for nasdaq and 3% for NYSE. On news days you can double or more these figures. You will always get differences between the yahoo open and the open cross for reasons previously explained. It is unreasonable to expect to trade at the yahoo open as this figure usually represents trades occuring at wide bid/asks in pre-open trading. If you want to measure how much you are impacting the open price with your order then I suggest comparing to the price at 09:35 for nasdaq and 5 minutes after whenever the nyse cross occurs. I would guess a figure of 0.25% shouldn't move the price much more than 0.1% though this will depend on when you are trading - flat open vs big gap up/down, news etc.
sam13, what's your experience with routing NYSE listings to NSDQ. Looking at the opening cross volumes (via ftp://ftp.nasdaqtrader.com/Files/crosses/) it seems that any significant order would result in a lot of slippage when trading anything other than the 50 most liquid, although I'm not aware by how much a large order imbalance would be offset.
I always route NYSE MOO orders to NYSE so no experience. There is good volume going through the NYSE, especially on news days.
Hi Sam this is a fascinating discussion. I originally sought answers to these questions many months ago (in another thread but also in this one) and I'm very pleased that you stumbled upon this thread. Can we look at a specific example? On 9/11 I shorted 500 shares of BAP on the NYSE. I did it with MOO. The trade executed at 09:36:30 @ 72.489998. The daily volume that day was 307k and the opening cross occurred at 9:36 with a volume of 3,700. I've attached an image of the chart. My 500 shares was 13.5% and the opening cross was 1.2% of the daily volume. I'm curious if I moved the price down. My trade just happened to be at the very low of the opening cross. I'm curious how others were able to trade in the range 72.49 to 73.48. It seems that for a short I got the worse fill possible. I'm also curious if I my strategy could be more optimal by avoiding the opening and placing the trade later. I'm not sure because the opening cross has the most volume of the 1 minute bars. But this isn't the first time I get a bad fill. I hope to do some backtesting on this once I get familiar with the backtesting platforms. I'm trying to evaluate them now. (chart in next post)