Appreciate the comments b.o.g...not refuting anything you're saying but the skeptic in me thinks this feels like a big curve fitting exercise. My two responses below are what would nag at me at if considering adding something like this to my set of models. Sounds spurious and unpredictable. Exactly...screw around with pairings and ratios long enough and you will certainly find perfectly mean reverting and/or trending combinations - historically. Without an intuitive reason for why that existed/should continue, it doesn't pass the sniff test (for me). In my trading, it's rare to have anything work out of sample unless I have a good reason for the anomaly to begin with. I'm not knocking bone's services at all - just some concerns I'd have about any non-arbitrage approach to pairs. In fact, if he has good answers to the above, I'd consider contacting him.
bone, can you help me understand this spread chart? http://elitetrader.com/vb/showthread.php?t=213015&pp=6&highlight=spread+chart&page=93
I use TT and have the Pro License by virtue of having license keys to several exchanges. Having said that, I rarely use the Auto Spreader and I do not day trade. We use the exchange spreads whenever and wherever possible. But of course, if I'm trading a synthetic inter market spread like Schatz vs mid curve Euribor or the Cattle Crush you have no choice but to leg it or use the Auto Spreader if you have it. It used to be relatively easy to day trade spreads - and indeed, I did it in Chicago for about ten years. These days, I think it wiser to take advantage of the margin credits and position trade spreads. I like for my clients to carry several diversified spread positions simultaneously. The volatility of the spread determines the trade holding lifespan - I might hold a Heating Oil Crack for 2 days, and I might hold a Eurodollar butterfly for six months. I don't think there's much sense in day trading spreads these days. I use the cheaper capitalization requirements to my advantage and swing trade them.
well thats why i posted the chart for you, that a strategy i trade with real money. sometimes i leg in and out of the components of the custom spread. but when it trends nicely, i can get in and out - take my money and go home. why does it happen? i dont know but i dont care, but it does exist.
on your website, it says you use the autospreader below for that strategy, i didnt realize the PJM electricity was available on TT.....i thought those were all clearport and OTC products. Above is my personal work station setup using the Trading Technologies Auto Spreader for the PJM electricity versus natural gas heat rate spread OTC swap. I can always get a better price here on the screen than through an OTC voice broker because I am picking up a considerable edge in either the power or the nat gas side. In the OTC world, the bid/ask spread can be a few to several thousand dollars or more for the smallest contract size.
Arbitrage is great - I'm all for it. Two problems: 1. The obvious ones that you are likely to have access to are way crowded and too automated. You will frequently get hung and flipped. Ex: Brokertec or Cantor vs CBOT basis; Comex Gold vs LME Gold 2. True arbs that are still profitable these days require access to OTC or physical markets - and that domain requires big capital and usually bonafides with the OTC brokers.
1.) there is no lme vs comex gold aribtrage.....there is an comex vs lme copper arbitrage and that trade still exists 2.) my point was that you said on your website that you spreaded PJM electricity vs ng heat rate swap. Those are not futures products available on TT....those are voice broker products... so how can you market that you are using autospreader for products that are not even listed futures?
There's a swaps tab in TT market explorer if you're enabled and set up by your clearing firm to trade Clearport and ICE-LCH. The market is more liquid and more products are available via OTC broker. It requires really big $$$ either way.
I think we have dramatically different styles. I trade based on statistical tendencies using very large sample sizes. I honestly have no idea what that chart was supposed to convey - looks like random periods of chop and trend to me. Even if the entire chart was a 45 degree line like the one atticus posted, a single chart wouldn't prove anything in my view of the markets. Again, not trying to be argumentative but rather make the point that you and I are unlikely to convince each other of anything since our styles are so different. Glad that combo works for you - this just doesn't appear to be a good fit for me.