I'm interested in the variety of CFD "prop" firms that charge an enrollment fee and challenge. I like the idea of paying a fixed fee to prove things out, perhaps primarily to myself. Are any CFD prop firms better at integrating with algorithmic trading than others? I mean beyond MT EAs: something where I can use my external data sources, generate signals and trade params R, and then hand off bracket orders to MT or another platform for execution? I know it sounds like a kludge, but I think it's the only way to get to what I want in terms of flexibility. Understanding the disadvantages of CFDs being less transparent and regulated, I'm also interested in what the transaction cost differences are - I know it's a huge success factor. Here's an attempt at a comparison trading between a CFD broker's average spreads (0.4) during a liquid market (i.e. https://www.globalprime.com/trading-conditions/spreads/) and IB during a liquid market (.25 pt spread), using S&P500 index (CFD US500, and CME MES): Let's say: At t1: S&P500 index is 4421.50 MES is trading 4413.00x4413.25 US500 CFD is trading 4421.30x4421.70 Let's model a winning intraday trade where I buy at market, the index goes up 2 points, and then I sell at market. The new quoted market would be something like this: At t2: S&P500 index 4423.50 MES 4415.00x4415.25 US500 CFD 4423.30x4423.70 I buy 1 contract of MES at 4413.25 then sell at 4415.00 I profit 1.75pts * $5 = $8.75 - $.50 rt commissions = $8.25 Margin requirement is $1100 Return on margin = 0.75% I buy 1 lot of US500 CFD at 4421.70 then sell at 4423.30 I profit 1.6pts * $1 - (zero commission) = $1.60 Margin requirement at 1:100 is $44.22 Return on margin = 3.62% Insane CFD leverage is alluring, but I think in most cases that is a fool's game. Let's assume the bid/ask midpoint is 'real' value - then what proportion of profit is given up (half spread when buying, half spread when selling, plus any commission) for transaction costs in order to participate in a given change in the underlying? For MES, we pay $0.50 in rt commission + .25*$5 spread = $1.75, or 17.5% of what a no-cost 2pt move would gain ($10) For US500 CFD, we pay $0.40 in spread, or 20% of what a no-cost transaction would gain ($2) It's actually not too bad - I was expecting CFDs might fare much worse. Of course, this is assuming those average spreads hold and there aren't a bunch of other shenanigans with CFDs. I created a table showing how, given all these assumptions, txn costs would impact profitability, required win rates, and net profits. Any feedback is most appreciated!
Nice post, it is not all black and white like many think. One nice advantage of CFDs (not with all brokers, but many) is that your max loss is limited to the cash in your account. You have 5k in your CFD account, and you can be sure thats the max amount you have to worry about. With futures it is different. With 500 USD intraday margins, it is no problem to trade 8 ES with a 5k account. With the ES at 4411, you are moving a value of 8*50*4411= 1.7 million USD. If the market makes a sudden drop of lets say 3% (for whatever crazy reason), you not only lost your 5k account balance, but your account will show about -45k. And the broker will for sure want to collect it from you.
Yes, this has to be 100% true. Even Interactive Brokes is offering CFDs to their customers, and everybody knows that they are the worst. Who are those few crazy people, desperados that have their accounts with them ?
@caementarius You also need to factor in financing charges applicable to CFD trades held overnight in order to make an accurate comparison.