Thank you. If your edge is in the "scalper" which has embedded "strict stop loss", I don't really see the value of the straddle. Do the majority of the profits actually come from the scalper or from the straddles? I have been pondering on a similar idea but in a reverse way ... Buy a strangle around a consolidation channel and run a mean reversion strategy for the purpose of paying for time decay. Trying to catch those fat tails on the cheap
Ok, lot of interesting stuff in your question. "Trying to catch those fat tails on the cheap" is a sort of "lottery attitude" of those buying strangles/straddles. And we can't deny that lotteries are a lot popular, even though we all know too well that the chance to win is ridiculously small. Undeniably, many people need to feed their dream, that one day one of those tickets, each of which seems relatively inexpensive, will bring a large win (your "fat tail"). And, even if that does not happen, they had anyway a dream to carry on. And, sometimes, dreams are even more important than their actual realization. On the other hand, thinking rationally, it is obvious that merely buying strangles is not leading far. Same for the opposite strategy, of selling them. We could say, that one might choose between 2 different and somehow opposite ways to lose money. For what I have determined, in this approach the "edge" is not the 1-contract "scalper", which by itself would be an unbearable proposition, nor the strangle. It is, instead, the precise combination: scalper (with a specific algo, of course) + strangle (repeated on the instrument price range) which is expected, in the long term, to converge to a profit. As you very intelligently noticed there are 2 ways, both legit, to look at this combination. You can think of a scalper "protected" by a strangle, or conversely, you may think of a strangle "financed" by the scalper. Whatever perspective you take, it does not really matter, because it is precisely the combination of the two the only one which has hope to generate your edge. As I mentioned in my first post, I duplicated the strangle, and the reason is that I was expecting that the combination of scalper + 1 strangle, should generate, in statistical terms, enough edge, to also buy a "lottery ticket" (the second strangle). This in recognition that the psychological aspect is also important for a trader, and we do need to be able to dream.
Here is an update about the "scalping" activity so far, and the global PNL: Global PNL: Margins: FullInitMarginReq 5,494.10 USD [Max: 8,411.90] FullMaintMarginReq 4,395.28 USD [Max: 6,729.52]
Update. Not much happening, the "scalper" is basically waiting for the price to move in the right spot to resume activity. (Options quotes have large fluctuations.) PNL curve (including closing cost): Margins: FullInitMarginReq 7,830.29 USD [Max: 8,411.90] FullMaintMarginReq 6,264.23 USD [Max: 6,729.52]
Thats cool stuff. Yeah, you actually might have picked the worse time possible to run the experiment.... strictly because of the geopolitical events regarding Iran and Venezuela. CL is always susceptible to news of that nature of course, but I think right now it is at an extremely heightened level. I bet your system will work great when (if) things calm down. Actually, you write quite well fwiw. Well said and so true. Great line.
Thank you vanzandt, very kind of you! Better so! And, if we start looking at the news, probably every time is going to look like the worst . A statistical algo, "unpredictive" by nature, is "agnostic" of whatever "market conditions", as perceived by us humans, and everything has to contribute to the big picture, for the stats to materialize.
Update. No trades in the last days, since CL was not moving. A couple of orders so far today. Margins: FullInitMarginReq 5,793.23 USD [Max: 8,411.90] FullMaintMarginReq 4,634.58 USD [Max: 6,729.52]
Update. Continuing on building up the "cloud" of orders. I have rolled over to CLN9 (CL FUT 201907 NYMEX 1000 Light Sweet Crude Oil). Margins: FullInitMarginReq: 5,957.33 USD [Max: 8,411.90] FullMaintMarginReq: 4,765.86 USD [Max: 6,729.52] (IB account is actually currently indicating a profit of 2,718.83 USD, but we incorporate the close expenses in our PNL computations.) Note that in our case, where we apply a strategy using an well defined algorithm, the "rollover" does not simply mean recreating the position on the new contract, but actually we need to transfer all the trading info. So we should rather call it an "information transfer". This is necessary because, for each action, the entire structure of the order cloud is taken into account.
Update. CL has been going down quite a bit, and the PUT option position (chart on the left) has been protecting the "scalper" tapering the drawdown. Margins: FullInitMarginReq 5,269.24 USD [Max: 8,411.90] FullMaintMarginReq 4,215.39 USD [Max: 6,729.52]
Update. CL has been bouncing up after the large move down, and I have just rolled over ("information transfer") to the new contract [ CLQ9 ]. The price did not go down enough to start a new "scalping unit" (we said we would wait beyond a 20% move of the option underlying). So we are still with the basic original setup. Margins: FullInitMarginReq 4,593.52 USD [Max: 8,411.90] FullMaintMarginReq 3,674.82 USD [Max: 6,729.52]