I'm looking to get into this too. I think the main thing to keep in mind is, Amaranth collapsed doing this.
I very highly doubt it, they trade like outrights. NG has far more mean reversion. at least with NG there are two definite ways to trade. Normal averaging in, and then when things move really far, ride the liquidation trend I guess. Note that I only started trading it yesterday. am play extremely cautious. This market is called the widow maker for a reason
Much more viable with Member Rates, consider leasing a Nymex seat for about $1100 per month. Problem with the strategy aside from slippage is doinking along for several tics life seems peachy and then getting your ass run over on sustained order flow in one direction - and since you are 'making markets', you are not pleased about it. Mean reversion generally speaking has not been a viable strategy past few years in a number of different markets - there are some exceptions, but rare. Flys and Condors more CTD friendly, but slippage a very big issue there.
Even with member rates its a bad idea in the crude spreads. (I've tried it) Bone is absolutely correct- at some points you get sustained order flow and you'll get destroyed (especially if you are trading size)
Yes even not during the inventories. When crude gets trashed quickly the front spread usually gets hit hard as well. This usually happens a few times a day- take a look at the charts. Many times mean reversion will work, but how much pain can you endure watching it?
Exactly. Market making is pretty much 2 things: (1) Staying market neutral... And any reasonable hedging strategy will be over the long run. (2) Capturing part of the spread with HFT. I've been doing this with stocks for 15 years... And have been experimenting with Forex and CL on and off... When I scalp something like CLM1... You got 2-3 tick spreads during the day, 4-5 tick off hours... Lots of things to hedge with... And you're not at the market... With a 2 tick spread you are scalping for 5 ticks, etc... A stone's throw outside the market.... After 10,000 learning curve trades it looks profitable... But still not sure if my results are statistically significant... A professional HFT operation should be able to grind out steady profits. What am I missing before I waste a lot of resources here?