I recently heard that the CBOT was looking into changing their order preference to go by size instead of by time. So basically if you were the first order in the book with a 5 lot, if someone comes in with 500 lot order behind you, he would get filled before you because the orders filled would be based on size not time. So the orders filled cater to the bigger traders. Has anyone else heard about this or know if it is true or not?
This is called a pro-rata algorithm and is used already by Liffe (and other exchanges) I have no idea if/when CBOT will change their current algorithm
I am familiar with the LIFFE exchange protocol but was surprised to hear the CBOT might go this way too. I am not a big fan of the pro-rata fills.
Interesting, the only two products that are currently on the FIFO (time-based matching, or first-in-first-out) are the treasury futures and indices. I can't see indices going to pro-rata, since CME's ES / NQ / ER2 are all FIFO matching algorithm based. However, the CME EuroDollar futures is pro-rata matching algorithm based, so it is possible (guessing here) that eCBOT would move the treasury futures to a pro-rata based algorithm. This would obviously give large size traders a substantial advantage in fills. The current matching algorithms for eCBOT is at: http://www.cbot.com/cbot/pub/page/0,3181,2000,00.html
Haven't heard anything about this. The Fed funds and all calendar spreads are on pro-rata right now. This works in the short end markets because they are less volatile and traders have to show more size to get filled. I don't think it is needed in Treasuries as there is relatively more volatility and more locals in the market (especially bigger locals). These bigger locals will show 5000 plus on the bid or offer and paper has no trouble at all getting filled. On the other hand, it seems this would definitely hurt Autospreader locals who stack the book. It seems paper only markets like Eurodollars and Fed Funds are a better fit for the pro-rata fills.
I think you are right on the treasuries, it is just very liquid of a market to need a change to a pro-rata based algorithm. I have seen the Merc's Globex vs Pit breakouts, since I am not a Board member, I haven't researched into the eCBOT vs pit percentage break downs. As usual, the autospreaders that just pile orders into the book is pushing the limit of the processing engines, so I figure that a messaging policy similar to Globex's would be in the works soon.
You would say in 2006 they could afford to upgrade their hardware, add some processors and memory here and there..
About fills on the CBOT Treasuries, is anyone able to explain me what it happens when - for example - you have 150 lots on bid and at a certain time you see hits of 100s or 1000s that hit the bid but those contracts are barely executed? Today I saw more than 6000 lots order that hit the bit on ZB but those few contracts were not executed, apparently.
Those are time and sales for the reduced tick spread, because it is rollover. The 6000 is being executed in the quarter tick spread, not the outrights, but it prints the outright leg prices used for the spread. Put up both the Dec and March time and sales next to each other and it will print in both simultaneously if it is a spread trading.