Do any ideas, theories, rumors, inuendo, or heavens forbid, even facts - exist concerning situations that offer even a hint of Anything Tradable or Testable having to do with futures front month / back month relationship, during last ten, five, or less trading days of the front month? Not concerned with swinging 100 oil contracts at a time, but onesy twosey folks. Preferred would be some completed spreadsheet magic, but ideas for generating spreadsheet magic are much sought and a close runner-up. Links, scribbles, or 'crackpot ideas' are entirely acceptable for this one. front https://www.investopedia.com/terms/f/front-month-contract.asp back https://www.investopedia.com/terms/b/backmonths.asp roll-yield https://www.investopedia.com/terms/r/roll-yield.asp https://adamhgrimes.com/how-to-calculate-futures-rolls/ http://verniman.blogspot.com/2013/07/trade-market-value.html thx
If you want to pay a subscription fee to seasonalgo, you get a pretty extensive database of seasonal tendencies of pretty much any spread combination you can think of. Everything is charted, so you can observe the behavior of the spread as it nears the roll date. Also, I believe (but have not verified) that "Goldman roll arb" profits have deteriorated as large funds have jumped on this anomaly, so historical data may not carry forward to more recent behavior.
Things get more interesting if you can ignore the prompt month(s) The "roll" trade is what you are alluding to; much of it has to do with commercials rolling hedges to the next month as the calendar spread is the cheapest way to accomplish that objective with size.
The front-month future converges with the index at expiration. This convergence creates a divergence with the back-month futures.