CL commonly fills at +1 tick better than prestaged... but more commonly slips -2 to -20 cents on stops. Both are due to 100-block or 200-block orders hitting the tape during lulls and into news reactions. The bonus flls are nice, but over time you will lose much more to negative slippage than gain on positive slippage. One more fundamental reason why "scalping" CL is a fool's folly only
It seems to me that since you are using a non-native stop market order type that the order is on your broker's stop server and not globex. It should be possible for the bid-ask to move in either direction prior to your order hitting globex.
Technically, CME's market order is "Market with Protection", which is translated into a LMT order with price is set at current bid(ask) + 1/2 product's Non-Reviewable Trading Range (6pt for ES, 1pt for CL). So clearly, unless in the midst of a vertical price jump, the Market with Protection order gets executed at current bid(ask). It seems most people think that STOP orders reside on the price matching (transaction execution) server, and are triggered & executed synchronously with the transaction execution. I think they are probably not, meaning complex orders (STOP, MIN-QTY, HIDDEN-QTY) reside on one or several separate servers, which use the transaction information flow to trigger whichever orders and send them in their execution format (a LIMIT order) to the transaction execution server. Hence the possible occurrence of positive slippage, which would otherwise be impossible.
Yes I can see it from the pdf available now. But I am sure I saw (a year ago) another document which was much more detailed where all market orders in the end was said to be converted to limits... and I cant find it anymore. either they dont want to disclose that info or matching algorithms has changed...
I commonly get "positive" slippage in CL or 6E and once in awhile TF by one tick... never more. All three commonly slip "negative" against by several ticks. I don't recall ever having the ES fill at better than expected price on a stop, and very rarely slips against. So that pretty much tells me at cursory glance it has everything to do with available liquidity versus block orders sweeping the levels at any given time.
Future is trading at 1000.00 / 1000.25 Your Buy Stop is set to trigger by a trade at 1000.25. A trade occurs at the ask. 1000.25. Latency .... Orders move about. Market re-quotes at 999.00 / 999.25 Stop gets triggered from fill above. Latency .... Orders move about. Market re-quotes at 998.00 / 998.25 Your Buy stop converts to a Buy limit @ 998.25 Filled at 998.25. Positive slippage. ... The worse your latency, the more variation you'll get in positive / negative slippage. One solution is to use native orders where available.
Ah, ok. Well, I guess there has to be some priority arrangement for everything that gets natively triggered. Good idea, to look for the matching algo / FIFO etc docs.
Here is some info from IB website http://www.interactivebrokers.com/en/trading/usFuturesStopOrder.php Stop orders are not natively supported on CME/Globex. Interactive Brokers provides customers with simulated stop orders