I've been doing some calculations on this but there are a lot of unknowns.
If you assumed a win ratio of 60:40 and you always went for 20pip targets and 20 pip profits then although SB is more costly, you actually get more profit from it due to the hidden nature of the spreads. The only way you could calculate DMA to be more profitable is if you added the extra 2 pips to the target of every trade because you are essentially making it by using limit orders. I'm not sure this is something people would do, ie they targeted 20 pips before, why would they they continue to target 20 pips again? Also, is the unknown factor of whether the lack of spread in DMA might cause your win ratio to increase.
On a like for like scale, if you win 20 pips on SB and you win 20 pips on DMA, then you actually win more on the SB platform because you have paid commissions on the DMA platform, which need to be deducted.
I have attached a spreadsheet, which you can play about with and see if anything looks obviously incorrect. Of course, if you changed your target to 22 pips taking account of the spread then figures would be different but so might your win ratio - catch 22.
For example, stick the figures up to £50pp and 22pips target for the DMA, still showing SB as more profitable...albeit I haven;t added in that ferrari into the expenses section 🙂