Does anyone trade CME futures directly without going through the FCM's risk control servers first? What kind of risk control deal did the FCM (Advantage, Dorman, Phillip, etc) give you? If I understand it correctly, the only risk control the FCM can do is by monitoring your after-the-fact trades and account. If they don't like something you're doing, they send a message to the CME revoking/ending your connection to the CME. The CME then rejects every instruction you send them. Which FCM does what kind of risk control? Or maybe some of them do no real-time monitoring and you can end up with a $10 million negative balance? Is the risk control percentage of account balance lost, maximum order size, maximum position size, a specific margin dollar requirement for each instrument, etc? I'm trying to determine if the only advantage of doing direct access is latency reduction because risk control is done after the trade, not before the trade. Or is there also a risk control advantage of doing direct market access because the FCM gives you better terms than a retail trader gets. Is it possible to get a deal where I have no maximum order/position size, no margin requirements of any kind and I can trade as big as I want(not for overnight, of course) and the only risk control is let's say auto-liquidation if 75% of account balance is lost.
On the CME website look for iLink access. You get your own iLink ID to use from your FCM, not bypassing them. You need to be certified with the CME and test with your FCM. Wedbush requires a $2mm account to do this. Risk controls and credit are set by the FCM.
I bypass the FCM's risk control servers and send orders directly to the CME from my own server located in the CME datacenter, thereby saving a couple milliseconds of latency. That's the main advantage of direct access: latency reduction. But, do FCM's also give you a better risk control deal? For example, for regular trading hours do I have no margin requirements of any kind and I can trade as big as I want? If the risk control deal for a direct access account is exactly the same as for a regular retail account, then the only advantage direct access has is latency reduction. That's what I'm trying to find out, is there also a risk control advantage?
The risk controls are on the iLink connection but set by your FCM. https://www.cmegroup.com/globex/trade-on-cme-globex/risk-management-tools.html They have to set your credit limit, margin limit, order size limit, position size limit etc. No one is allowed unrestricted access without risk controls.
Really, the CME got in the business of calculating your account balance, margin requirements, order size limits, position size limits? That's some seriously bad news for Rithmic/CQG/TT because that's their business model - allowing FCM's to outsource their risk control so that the FCM doesn't have to operate their own risk control servers. So, I guess you're saying the FCM is not going to give you reduced margin requirements if you have a big account size and the only advantage of doing DMA is to save 2 milliseconds of latency.
ICC= Inline Credit Control. This is the CME risk system an FCM can enable at the account level. There is also GC2. This is at the Execution Firm level. Once your order hits the CME systems all things are equal. From the cable length to the risk calculations that are performed. All orders run through GC2 and ICC. The only difference is if enabled your order could reject. The basic path for all orders is CME public facing interface - GC2--ICC--CME Order Gateway.
Admittedly my understanding of this topic is several years old, but based on that knowledge... CME is counter-party to all trades. A clearing member/FCM has final determination of risk management etc associated with your account. YOU are free to become a clearing member/FCM if you qualify and so choose to do so. As an independent trader, you are free to negotiate with clearing members/FCMs of your choosing before depositing your funds. Whether or not the FCM "operates" their own servers or merely provides instruction and/or parameters to the counter-party is not in question. EDIT: Thank you for the quick lesson Matt. Just saw your post.
That really doesn't make any sense at all from an FCM's risk control perspective (or for a fund/firm's internal risk control aspect for that matter). Most sophisticated (bigger) traders and fund/firm accounts have on simultaneous futures positions from multiple trading exchanges - not just the CME. There's also some SPAN cross-margin and offsets going on for spread trades and for hedging purposes.
There was a time that it made sense for a pro or a prop desk to have CME Direct because it did not charge fees for market data. Now they do, so I do not see any advantage for having CME direct over CG, TT or Rithmic when it comes to latency, execution or data. Fewer and fewer traders are looking to trade through that directly, and when they really want to improve latency they use a direct API programming along with good hosting.