I don't think you got the gist of my remark. I never questioned your results. I questioned the honesty and competency of the rest of the Wall Street, other than IB. Also you didn't address the psychological benefit of not having to endure a given expense of 50k to 100k. To me paying that amount of money is liking paying dues for a club memberships and concluding that the benefits of membership is worth more than that amount. My conclusion is that it is not for reasons given.
The cost savings I referred to were net of all commissions and exchange fees. I agreed with you that questioning the honesty of parties you interact with on Wall Street is important. That's what prompted me to put them to the test regarding my order fills. If you are having a hard time with psychology then I'm not really the guy to help you. All I can say is that adopting an objective thought process is helpful. The intentions of all the parties involved in a trade are not nearly as important to me as the end result.
Your methodology involves error an extrapolation in drawing a conclusion. Your sample of two is too small when considering the knowns of how Wall Street operate. Firms can tell me that they are saving me a given amount while paying a commission, but except for IB I will continue to doubt them
Interesting study that you did. Did you do any study of non-marketable limit orders, IB vs. "free" others, by any chance?
I know this is an older thread; but may I ask what kind of equities you used for your test? I'm asking because my long-run tests show the opposite result - almost always better execution at Schwab (formerly TD Ameritrade) than at IB. Also, how does your result reconcile with the study "The 'Actual Retail Price' of Equity Trades" https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4189239 that also shows the opposite? Lastly, you are saying "the cost savings I referred to were net of all commissions and exchange fees". So you must have used less liquid equities with more than $0.01 bid/ask spread; otherwise the roundtrip commission and fees alone (ca. 2 x $0.0067 per share for market orders at IB) would be higher than any potential savings from price improvements, correct? I hope you can clarify how you set up your experiment.