I write all quote data from TWS to disk by symbol... So I collected a week x 24 hours of several CL contracts... Then wrote a VB6 program to parse the data files. The numbers are entirely reasonable... Last week I estimated the June CL at 2 tick ave spread. But volatility widens the spread... So it actually averages out to 2.5 ticks... The near March contract is < 1.5 ticks average spread. In off hours the spreads may be 4-5 ticks... But it's hard to make a trade... Though every time you do... it's at a premium.
You can't scalp those back months like that, because those are markets being quoted (almost always by automation) based off the front month. You will not even be able to get filled in the back month and then work the front month as a legged spread. That "scalp" fill in the back will be you getting picked off - there are guys with dedicated lines ($$) and co-located servers who will pick you off every time. I personally know of a team who located across the street from the ICE server farm in Chicago and six blocks over from the CME in order to do the ICE WTI vs. Nymex WTI arbitrage - and he still got picked off and had to give it up.
Give it to me, man... This is exactly the kind of stuff I wanna hear... So I don't waste my resources. But I've made 2 million stock trades over 15 years... Because I understand market randomness. "back month... based off front month" Yeah, they only diverged by 70 ticks or whatever Friday... Going back and forth all week long... And you would rarely be hedging that way... Because then your business becomes trading Calendar Spreads... Which is all about fundamentals like geographic supply... Or outright manipulation like Brent right now... Not capturing the 2.5 tick back spread. "fill in the back will be you getting picked off" You are assuming I'm near the NBBO or whatever it's called in futures... And moving with it. No, I'm always 2-3 ticks OFF the NBBO... So my fills are primarily the result of random movement... I'm sure I will get "picked off" and pay on 10% of my fills... But a 2.5 tick spread can absorb a lot of misadventure. I would NEVER go anywhere near a strategy... That pits me directly against 1 ms colocated Players... That would be like walking into the Bellagio... And sitting in the Big Game. I simply wanna pit Random Movement that NO ONE can predict... Versus a 2.5 tick spread... and stay out of trouble. And I don't believe this (or pretty much anything else)... Can be accurately "back-tested"... You will rarely be able to simulate actual market conditions. You gotta make 10,000 trades and reevaluate... then tweak or quit... I got my core Stock Market Making business anyway... This is a "fun" way to upgrade my technology across the board.
DD2, I am personally very active in the back months for most of the serial (12 month) intracommodity spreads like STIRS, Softs, and Energy. But I am trading Flys and Condors and intercommodity relative value spread positions with holding periods of days (and longer) in order to capture divergence or convergence. When I execute them, I am doing combinations of exchange-supported implied spreads. Given your business model, and it is a good one, IMO there are better places to invest your resources. Just my 2 cents. As a side note, it is going to get tougher and tougher for you to find located random walks in defined trading ranges. I have done some consulting work for a couple stat arb firms past 18 months, but NDA will not permit me to elaborate - wish I could.
You're doing stuff that requires a 10 year learning curve... Very inside baseball... But you MUST be capturing the spread on your option positions... Or it's all a non-starter regardless of your expertise. Providing liquidity in return for the spread is a core Business Model... But it's a Zero Sum Game, so 95% fail to EXECUTE... And we are not talking latency here... Since I view the 1 ms playing field as a SEPARATE Zero Sum Game... With the same 95% percentage losers despite the big tech budgets... With the vendors = big winners. I've done enough detailed analysis to see opportunity in a 2.5 tick spread... You pile up the ticks on sideways days like Thursday... And find CREATIVE ways to limit the damage on big move days like Friday... I should have little Algo Bots criss-crossing CLM1 by Sunday night. Probably should find the time to read your last 100 posts