Wait… What? Fred and Ethel Have a Smart Order Router?

 

 

It is hard to believe that it has been over two years since retirees Fred and Ethel first contacted us to learn about modern markets, and how they work. In those two years, our discussions have run the gamut, and have included detailed sit-downs where we have discussed Fragmentation, Sub-penny Internalization, Intermarket Sweep Orders, and even IOIs and Dark Pools. The quickness with which they picked up the subject matter shocked us, and we knew very well that They Had Game.

However, nothing could have prepared us for how well these students became the teachers. They visited with us yesterday, and when we asked what they wanted to learn about, Fred quickly chimed in.

 

F:         Oh, we are not here to burden you. We want to offer you something!

E:         Yes – We have been really digging into how our markets work, and have uncovered some problems that need a fixing! Tell ‘em Fred!

 F:         Jesus woman… we discussed this, remember? Whet appetite, give a little smidgeon, then bammm hook! Stick to the game plan! Anyways boys, it goes like this. I know you think probably that we ain’t connected, but that there is not true. Ethel and I actually sat down with an old army buddy of mine – Charlie Munger. Maybe you heard of him?

 S&J: You guys are connected to Berkshire Hathaway? Well I’ll be a… wait you want to offer us something?

 E:         That’s right gents. It goes like this. We analyzed the trades that Charlie and his company did. We got data with timestamps and everything! You won’t believe it… he has this trading system that gives him every execution real time, and all the information that surrounds it.

 S&J:    Like what kind of information? You mean like an average price for many shares?

 F:         Let her finish boys. No, not average prices… We got data for things called child orders – each individual execution. We got who did the trade, and we even got this 30-tag which tells Charlie where the trade was executed.

S&J:    That’s great guys! But do you even know what to do with that information?

 E:         That ain’t all! We got other information too. We got this tag that lets you know whether the trade was executed as Principal or Agent, and we got route data!

S&J:    Wait… Charlie has route data? He knows where each order was routed by his broker, whether it got executed or filled or not? He knows where each order went before it ended up finally getting a fill?

 F:         That’s right tough guys! And we got all that data. And we got like a DVR of what was going on in the market place at the time. Here’s the thing – Lots of Charlie’s trades got broken up into small trades, and even the small trades traveled from Michigan to Chicago by way of San Diego.

S&J:    Ahh You mean you guys were able to see the data as it relates to the Sausage Fac..

 F:         Don’t cut me off. I’m saying that even to execute an 800 shares trade, that was part of a large algomaslicy thing, the 800 shares would not just go to an exchange directly. It would first go sit in the algo owner’s dark pool for a spell – I forget the name, but it always happened first-off. It would always go into the owner’s dark pool even though it only got a fill there about 5% of the time, and then it would go to these firms I never heard of…

S&J:    Ahh you mean ELPs – “electronic liquidity providers”

E:         Exactly. But most of the time they were shown to these ELPs, they would not execute the trade! They would only trade about 5% of the orders that they were shown also!

 F:         Yup. Ethel is right. And then they would go through a bunch of those other dark pools you taught us about. And if they couldn’t find a trade there, they would get spit out to the stock exchanges – and not the one that is on TV with the cameras and that Eyetalian girl. And when the orders finally went to the high-rebate exchange without the Eyetalian girl, it tried to post a bid, and then a higher bid, and then a higher bid… even though it would have been better off to just take the first offer.

E:         Yes… because it wanted to keep trying to get a rebate. One thing we both learned from real estate 101 is that the first offer is usually the best offer!

S&J:    Guys guys… slow down. We know this already. It’s called the Sausage Factory. We have been though this with you.

F:         Not completely, you haven’t… You told us that this thing called a Smart Order Router would route Charlie’s trades in a smart way! Hog wash. We actually called the broker whose algomaslicy thing Charlie uses… We asked them how they choose where to send the 800 share slices, and with what order types. THEY told us that Slicy looked at the broker’s historical database, which looked at where would be the best place to get the trade, and in a smart way optimized everything so Charlie would never have to worry – spreads being a penny wide and all.

 E:         Not only that, the broker gave Charlie this pretty report each month which showed how much of his trades happened in between the spread, and how much money he saved.

S&J:    Yeah we are not a fan of those reports guys. The spread is meaningless because it’s …

 F:         Of course it’s meaningless numnuts! Stop cutting us off. We know the report is horse manure. I said to Ethel, if Slicy were so smart, and changing on the fly, why did it always seem to go the same way, time after time? Why does it always rest first in Slicy’s pool, then go to ELPs, then go to dark pools, then to exchanges finally? That made no sense! Why would Charlie want all those people knowing his business? Also, Slicy sent out lots of IOI pingies every chance to those ELP guys who traded it only 5% of the time. Only the Lord knows what those ELPs did with the information the other 95% of the time.

 E:         And then I got to thinking, there has to be a reason why Slicy does it that way. I looked at each route – each pitstop, and what was happening to the stock during the period, and I looked what happened when Charlie finally got a trade done, and what happened to the stock in the next minute! I also looked at what happened to the stock after an ELP got the supposedly blind pingy. It took weeks, and I went through six batteries on my calculator, but I finally figured out that this was really costing Charlie real money!

S&J:    Yes guys, but doesn’t Charlie also get a report that tells him how great he bought versus things like the VWAP?

 F:         I looked into that VWAP thing. If there ever was a dumb thing to compare to… It’s like comparing yourself to Dan Pasqua.

 S&J:    Who the heck is Dan Pasqua?

E:         Precisely boys… he played for the Yankees and the White Sox. Batted about 240 w/ 30RBIs a season.

 S&J:    Ahhh We understand. So are you saying that the pattern of routing and executions by Slicy looked ok compared to VWAP, or Dan Pasqua, but when you looked at each fill, and each route, and what that did to the stock – and the few pennies it moved it, and then reverted, it looked less good?

 F:         Bingo boys! And we dug into why Slicy really did what it did! Historical smart database my ass – it was trying to favor its own dark pool, and then it was trying to get paid by ELPs who paid them rebates to show them the orders even though they only executed them like 5% of the time, and then through other dark pools… and it was trying hope upon hope to get rebates finally on exchanges. Heck, even a lot of the time that the order was even executed by the Slicy’s pool – supposedly in a clean way – it was a Principal transaction and not agency! I mean the order was like that Johnny Cash song – Everywhere Man.

 S&J:    Wait! We know that song!: https://soundcloud.com/algosalesman/everywhere-man

E:         That’s the one boys! And that brings us to why we are here today! We were telling the story about Charlie’s trades to this nice young man we met a Dallas Mavericks game – Mark Cuban, and he asked us why we didn’t’ create a new Slicy that didn’t beat Charlie up so bad.

 F:         Yes indeedy… and this Cuban fellow gave us some money too! Ethel and I have created a Smart Order Router called CommonSensy. We want to show it to you today!

S&J:    No way! Tell us about it!

 E:         Well, Fred and I realized that even a random router that sprayed to destinations that didn’t offer rebates would probably be so much better than that Slicy.  We started with that premise, and built on that with real smarts. Just by taking out dirty low-fill % pools, and people that pay rebates, we would have a router that just HAD to be better than Slicy.

S&J:    Wait, you are pitching us to use your SOR that costs more than Slicy?

 E:         You betcha! It costs just a little more, but it saves you ten times that by not checking you against the glass like a hockey enforcer every time you trade.

 F:         Wait guys it gets better. Let me tell you more about CommonSensy. It starts out unconflicted like Ethel says, but it gets better. First off it uses a data base that is set up in the cloud for historical execution information – 3 months-worth. Like for example:

–          where stuff actually got filled,

–          what were the fill rate percentages,

–          and what happened after the fill 5, 30, and 300 seconds later.

 

Secondly, it uses a second data base that starts fresh every day. It remembers where the trader executes, and weights very heavily recent memory. If the algo gets fills in IEX, it sends to IEX first and larger each time, each child,… and it ups the size and will always weight that higher than the 3-month historical data. Why would anyone want to expose their order to more places than they have to, especially if they found a pool with clean water? By the way CommonSensy goes to about half the destinations than Charlie’s Slicy did. After all, there are some bars one ought never to walk into in the first place. My father taught me that.

Third, it never goes anywhere with a low fill rate, and as Ethel said earlier,  it never goes anywhere that pays for it to go there – i.e. we get paid no rebates from anyone. Why, that would be meshugah!

CommonSensy executes you in this fragmented mess of a market by only using clean destinations in an unconflicted way. It also decides where you go by weighing a recent day twice as heavily as what happened in history. And… if stocks don’t execute cleanly in the few dark pools we go to, then it goes to the exchanges exactly like SensyLite does.

 

S&J:    Guys, this sounds just like pure common sense and not that complicated. Why does it cost more than Charlie’s broker’s Slicy?

F&E:   Simple. CommonSensy gets no rebates, and incurs take fees more often. And the proof is in the pudding. Charlie has tried it out and he gets better prices and impacts on average about 8 cents better on the average price than Slicy got.

S&J:    8 cents? That’s huge!!!

F&E:   We think so… Charlie buys 200,000 shares of stock 8 cents cheaper with our bad boy than Slicy!

S&J:    Wow guys. Who wouldn’t pay pay a penny per share instead of half a penny if it saves you 8 pennies! Can we try it out? Test it? I think the buy side might take a while to warm up to actually paying more for an algo, but I also know that their frustration with shoddy execution practices and results is high.

F&E:   Then why do some of them say that transaction costs have never been cheaper?

S&J:    You know the answer guys. The measures most commonly used are things like VWAP and spreads. One is a self-fulfilling target that can vary wildly depending even on how dirty one VWAPs, and the other is meaningless today.

By the way – and we are just thinking big picture here – if your router works the way you say it does – there would be other less-tangible benefits too!

F&E:   What do you mean?

S&J:    First off, if it hit less dark pools, and other firms changed their Slicys to act more cleanly like CommonSensy, then some bad dark pools would fold. There would be less fragmentation. Also, I think there might be more institutional order flow diverted back to stock exchanges too! This means that public limit order books wouldn’t just have HFT guys like the really insanely profitable one going public right now; the exchanges would have diverse players – and be safer in times of stress!

F:         I see your point. It’s amazing how much cleaner the market can get if you just take out payment for order flow. So, will you try it out? You think it will catch on? What can we do to close the sale? By the way Ethel brought over her squirrel stew. Where are your bowls?

S&J:    Squirrel stew? Ummmmmmm…