I don't rule out the possibility of improving execution by judicious use of limit orders. For example, the spread can sometimes be too large to pay.


Also, execution via limit orders can be superior to market orders if you can place your limit orders with a brokerage firm that does a lot of internal order matching in the stock you're trading and a large percentage of its market orders (in that stock) are placed by naive (rather than sharp) clients.


Here's an example (that's probably more favorable than what you can hope for): back when NASDAQ still quoted in fractions, while some ECNs quoted in decimals, I used limit orders to trade one particular stock with Brown & Co. and got tons of fills at the NASDAQ quote that were outside the best bid and ask available on the ECNs: the NASDAQ spread couldn't get tighter than 1/32 while the ECN spread was typically one cent and sometimes less. For example the best ECN quote might be .31 x 5/16 and I'd fill my buy order at 9/32 or the quote might be 9/32 x .29 and I'd fill my sell order at 5/16. With fills like these, I accumulated a ton of shares at a negative cost.


... Note that this also means that Brown & Co. didn't give its clients the best possible execution on their market orders. This is something you must be wary of when you're selecting a broker. It's possible that this is less of a problem now that NASDAQ quotes in pennies. Then again, some ECNS accept bids in mills so your broker could still fill your order at 0.9 cents worse than the best price available.