PDA

View Full Version : PROOF AT LAST! two ways...


11-30-2001, 08:05 PM
The first proof limit orders, in and of themselves, don't offer any benefit - and I don't know why I didn't think of this - is that if they did, Paul would have played both sides and made free money. How obvious. Brown was handing out gifts and manufacturing millionaires.


If one is good, two is better - and Paul could have bid and offered a penny outside Island and made free money all day. After all, that 340,000 he was leaning on is pretty big! And he was getting filled in the absence of any directional market bias or adverse selction. If his orders had negative cost, why not put in a hundred thousand of them (offsetting of course) and be a day-trader?


Of course, what this proves is not that limit orders win, but that asymmetric friction, even coupled with symmetric or inferior access and information, wins. And the reason why is that limit orders - in proportion to friction - lose. In reality, your fills will be skewed, where you'll hit on both your limit bid and offer 5 times - and make 5 * 3/8 - and then it will hit your bid and drop 2 points like a stone before lifting enough to hit your offer and offset.


The second proof limit orders don't offer any value is a little more indirect, but also more applicable to Ray's discussion. Suppose the stock you want to invest in is bid-asked 20 x 21. So, you hope to buy it for 21, and sell it for 40 next year. But by bidding for 20 1/2, you think you can squeeze an extra half point out of it today.


This is very clear, you hope to tack an extra half point onto your long-term profit by making a half point today. Put differently, you hope to buy at 20 1/2 as a day trader, and then sell to yourself at 21 as an investor. No matter how you look at it, you're saying that by buying at 20 1/2, you can pull an extra half point out of it today.


But the market-maker bidding at 20 has more infromation than you, and he doesn't think he can buy at 20 1/2 and make money today. He thinks he can either buy there, or theoretically make money there but not get filled. Of course he could make money if he bought it on Mars for 20 1/2, and then sold it here on Earth where the price either didn't budge or actually went higher. But in reality, he knows he could only buy at 20 1/2 if the market went down, 9 out of 10 times.


How could you be the only person on Earth who thinks your odds are good enough to actually buy at 20 1/2 other than in the moment 20 1/2 becomes "the offer" - and the bid drops to 19 1/2? The answer: It's an illusion. If the market-maker, with all types of advantages, can't make money bidding - not necessarily buying - at 20 1/2, then you can't either. And the only reason you think you can make money buying at 21 is because you - unlike the market-maker - have some reason to believe it will be at 40 in a year, and nothing better to do with your money.


If there were an extra half point plainly available, before the close, through some contorted bid-ask magic, the market-maker would take it himself. The half point isn't there.


loudmouth

11-30-2001, 09:20 PM
I think you may be right (in fact, I'm hoping you're right), but it's still not clear to me how I could have made more money. Maybe you can identify what I'm overlooking.


On Brown I was in fact simultaneously bidding and offering as tight as I could (i.e., on 32nds) outside the ECN quotes. I was making free money, but only when Brown matched one of my limit orders internally with a market order. My gross was 1/32nd (or better, occasionally) on a round trip, less 0.4 cents commission, so 2.725 cents per round trip.


You're saying that instead of this, I should have been bidding and offering a penny outside Island. Is this different from what I was doing?


As it turns out, it's possible that I did milk the situation for a good chunk of its potential. One cent profit on 300,000 shares is $3000 and I made more than this before the situation ended.


But if there's a way to do better than this, I am all ears!