I dabble around with butterflies a lot. It's my go to position for trading vol together with the risk reversal (which is basically just half a butterfly). When I trade vol in a single name I always have some sort of vega neutral fly i.e. more wings than body. Recently I was asking myself where the no arb price for the 25 delta/50 delta relationship is (aside from basic vertical no arb criteria of course), meaning how does a risk neutral skew look like? Usually I look at historical 25 delta butterfly prices to determine when wings are cheap, but I'm not quite satisfied. When I get a shitty execution, my vega neutral fly is short gamma and short theta (meaning wings are bought too rich). I had the idea of tracking the gamma/theta relationship of wings vs body to figure out where the no arb vols are. A well executed vega neutral fly has almost zero gamma and is long theta. I figured if I got a fly that has no gamma/theta/vega but is long vega convexity its basically a free trade. There is a ton of papers regarding arbitrage free volatility surfaces, but I don't find them very practical. Perhaps there is a better way.... P.S.: Yeah, it's very niche and no, it's not a trade for retail with 5$/contract commission, yes, you need a big account to get the ratios right.
I'm wondering if verticals with long strike at 15 deltas, ratioed to be neutral of gamma, theta, and vega while giving you maximum vomma, meet your criteria. I always wonder how to manage this sort of trade. It won't stay neutral for very long. What if volatility doesn't change as fast as price or time? The vomma won't produce before the trade has to be re-hedged. Admittedly, I've never tried this, exactly because I'm vague on the particulars, but it's on my wish list to do this somehow. Another question becomes how to find the right ratio. What software, if any, would you use to meet your outcomes? What underlying would you use for trades like these, or does it matter?
That's why a vertical is out of question, it needs to be a fly. Otherwise you would fall victimm to serious skew deltas. Also, you need to manage positions somewhat, especially when the underlying moves towards your long strikes. the right ratio is pretty easy. You need as many vegas on the wings as you have on your short strike. However, it's not easy to determine if skew is cheap or not.
I just put up a separate thread based on your post: Is skew cheap or expensive? I look at historical slope (our measure of skew), ratio of slope to the best related ETF, ratio of slope to the components of the best ETF, and ratio to the forecast slope.
I know everybody has different opinions but the papers about free arbitrage free volatility surfaces were very useful for me.