1 soybean vs 2 soyoil works for me.The bear trend for soybean is kind of obvious,so if not for strong crude oil,you don't need that many of long legs
I too would also guess 1 bean by 2 oil or 1 bean by 2 meal At my old firm we had an algo that used to trade meal vs. oil 1x1
So 3 years later I am running into this same problem. Anyone know what's up with CME's margin offsets for soybeans? They don't make any sense to me: What's up with these 5 and 50 numbers? That doesn't match up at all with what the normal crush ratios are? I bring this up because I was just academically checking out bean vs meal spreads and noticed that the ratios I came up with on my own and from CSI's site bear no resemblance to the reality CME seems to use for purposes of margin.
You could also use implied volatility (IV) which is a better at forecasting volatility than historical volatility. And then you try to get equlibrate the equation nb contracts ZL * ZL values * ZL IV = nb contracts ZM * ZM values * ZM IV Contrary to spreads, ratios are dynamics so you will have to readjust you quantities to get same risk on both legs Assuming that spreads are two products with the same number of tons.
The issue here is that the crush is well known to be 10 beans, 11 meal, 9 oil - even by CME's spec (http://www.cmegroup.com/trading/agr...ed/soybean-crush_contract_specifications.html), so why would 100% margin offset against the crush be 50 beans, 9 oil, and 11 meal? The same issue is seen with beans vs meal, which uses a 5:1 ratio. These numbers are not right.
I contacted CME risk management on this and they basically confirmed that this is a quirk in how their SPAN system feeds data to the site and indeed the contracts showing multiples of 5 are the mini contracts even though it appears as they are the full size symbols.
Most traders wouldn't consider contacting the exchange directly so I'm glad you did. I have called them a couple times over the years and have found them to be very responsive and glad to answer questions.