there's no offsetting long to this trade because the drop was due to an acquisition... so it's paired with SPY, which is the long. that's an important part of the the system rules.
Background info on earnings in general. This will be review for most readers: There are at least 4 important parts to monitor with an earnings announcement: -The revenue number (the "top line") -The EPS number ("bottom line") -Guidance -Company comments in conference call (may not be entirely consistent with guidance.) What really matters with these is how they relate to expectations. A company can report a massive loss, but if the Street was expecting an even worse loss, this can actually be good news and the stock can trade up. It's maybe a little counter intuitive, but from a traders perspective I don't think there is any real benefit in trying to understand what the numbers should be or trying to understand how they relate the the future prospects of the Company. All that really matters day to day is how earnings relate to expectations. it is also possible that a great earnings beat can be tempered by negative comments in the conference call. To further complicate things, guidance doesn't have to come with earnings but can come anytime and guidance can sometimes have a stronger effect than an actual earnings report because guidance is often followed by analysts' changes which leads to massive amounts of money flowing as the sell side sales desks get busy calling their clients. So... this isn't all that easy to quantify...
Talon, When you refer to "we" in your testing and trading, are you refering to a group of private traders, a firm, a fund, etc? Just curious
I shorted an equally small position of PCLN and went long SPY at market close last Thursday. I used MOC orders with IB. I think I will only use those as a last resort if I wont be at the computer to trade because I did not get very good fills. I am currently short PCLN @ 167.63 (-11.5%) and long SPY @ 106.96 (+2.51%).
I was using robust in a casual sense, meaning that the tendencies we find, and the models and systems created from those tendencies tend to work year after year without too much adjustment. Most quant models are not robust in that sense... they work for months or years and never work as well in actual practice as they did in development. (And when I say most quant models I'm not talking about most stuff on internet boards understand.) It's simply a different approach. Of course in our minds it's a better approach (or we wouldn't use it), but the idea is we think we have some ideas about the ways markets "should" work and we develop trading models from those ideas. To us, going to the data and blasting away with a massive multidimensional matrix of parameters until you find something that looks good and then applying a data mining correction (which btw, doesn't really work) is a crude and blunt instrument. I'm not saying you can't find trading edges like that, but I do believe that the ones we find are more durable (maybe that's a better word than robust) and ultimately we have more confidence to put money behind them. I have seen COUNTLESS models developed by brilliant people fail because of data mining, but if it's working for you then you must be doing something right.
1. You should be fine with MOC orders, but make sure they are NYSE MOC so you get settlement. I dont have numbers here in front of me at home but that PCLN price at least looks wrong. You want to be able to route your MOC to the Specialist. 2. Did you test this first? I would never trade anything just because someone told me it was good without doing due diligence!!!! You are ultimately responsible for your own results so you gotta do the work.
And what advice do you have on the relative size of the paired spy trade. I would say pair trade, but that is not what a pair trade is. So, if we were long 10% of our portfolio in PCLN, how large should SPY be in our portfolio? It's a very important question, if you ask me, and might be a determining factor in how robust your system is. ex post edit: Your system description answered this question. Equal dollars. So at 10% of equity you shorted PCLN at 166 now at 173.73 and bought SPY all on November 5th on the open at 105.66 now at 109.57. So .1*(166/173.73-1)+0.1*(109.57/105.66)= -.44494%+.37005%. Net off 7.489 basis points, or 0.0007489, or 0.07489%. On a $200,000 portfolio without transaction costs= ($149.77). Yeah.....We'll see what happens tomorrow and till your 25 day expiration.
I have a couple trading assistants / junior traders who are somewhere on the learning curve between slime mold and full time traders. (And yes they do read these posts. They need to be reminded of humility and the dangers of hubris after particulary good trading runs lol.) To me it's very much a collaborative process. This will probably start a flame war so I'll bury it here. Let me make a bold statement: the market is not a math problem. It's art masquerading as a math problem... the information the market gives us comes to us in numbers and quantifiable form, but the essence of the problem is not mathematical. I know there will be much debate and anger over this (no one gets more angry about this than engineers hoping to make it as traders, but I don't hire engineers... also don't hire accountants or analysts... but that's another story). We can debate this more in the future... but tonight I'm tired and out of energy.
Dollar neutral at inception. No adjustment for volatility or beta, if that's what you're asking. Hard to give advice on position sizing since risk is so undefined with this system. My sense is that if your entire trading capital was deployed in this idea that 10% would be FAR too much... but we have a certain % of capital deployed to this idea and then more than 10% allocated to each trade so maybe not. You just have to be aware of the possibility... which becomes a virtual certainty if you do this long enough... of a staggering loss on one of these shorts. You can find ways to deal with it, but our solution is just to trade small and sleep easy.
Thanks for the MOC tip. Not sure who the order was sent to as I used smart routing. I knew right away I had a bad fill but that happens. No biggy. I did test. I spent a few hours and went through all the trades in 2009 and late 2008 and liked what I saw. I am going to track and do this for the entire year and see what my results are. Just with small positions.