time value will decay faster on your long put. vega sensitivity will inflate the short rear month put if VIX spikes as well (since IV of vix will shoot up as well).
furthermore you have to be realistic about the difference between the months you expect this to get to. just looking quickly, it looks like +.25 is a realistic price target for the spread. look at the chart below. i think we're at somewhere between -1.00 and -.150 right now.
for a $1.00 profit (on the options spread you suggest) at most, this looks like an aweful lot of effort ... especially having to factor the complexities of the option spread.
And another problem with playing it with puts -- what if the calendar repairs itself (ie rear months go into contango) as you say, but vix stays slightly above 20? You lose all of your long premium, and have to likely close out your spread at breakeven at best (since that time decay kills your options spread).
playing the calendar on -sep/+oct here has had about $1500 of total risk per spread if you long it [with volatility blowing up]. If VIX blows up even more, you'll see even more risk in doing the futures spread. (another $1500? with similar magnitude adverse vix moves)
Maybe you can do something of a hybrid to hedge from blowing up, ie 1 -SEP/+OCT VIX futures spread, against ie 3 long SEP 25 calls. $750 for the calls will definitely cover your short SEP blowing up, but a positive calendar move for your spread will get you a $1500 gain (so total of $750 profit, if vix closes under 25).
And of course there is a *remote* possibility vix stays elevated and you get an expiration above 25, while you get the calendar move you anticipate.
