View Full Version : Theory Question

01-17-2002, 01:21 PM
Okay, suppose I only trade one market, so I am always either IN or OUT. And anytime I am in, I am either LONG or SHORT.

At occasional moments, I get either a BUY LONG or a SELL SHORT signal, meaning at those moments I have an expectation that the market will move enough up or down to cover a commission and a slippage.

Given that there are no other marketplaces competing for my margin dollar, but assuming I am averse to volatility, is it ever rational for me to be OUT?

Meaning, given that when the market goes random (a signal runs it course) I am left LONG - and therefore the next signal is just as likely to be BUY LONG or SELL SHORT - under what conditions would it be rational to exit a long position (and incur commission/slippage), but to not at the same time enter into a short position?

Meaning, if there is not a strong enough short signal to cover the entry/exit costs of a new short entry... well, please explain the situation:)


01-17-2002, 02:15 PM
I don't know if this answers your question since you like to theorize quite a bit but frequently the best trade is no trade.

01-17-2002, 02:23 PM

01-17-2002, 02:28 PM
"when it feels right" "NM"

01-17-2002, 02:29 PM

01-17-2002, 04:03 PM
I think we need to characterize the commission, slippage, and signals a little better, at which point we can just solve this puppy (although maybe it would be like "just solving" the wave equation, who knows).

For example, do we have more slippage if we trade a bigger block? Is the commission a fixed amount per block trade, or a fixed amount per share traded, or a combination of the two? Is the signal a "a position taken now is guaranteed to increase in value by 5%", or is it more complicated?

I think the simpler the signal is, the more likely the answer is going to be either "always get out when the signal disappears" or "never bother to get out", depending on the definition of commission and slippage. The more complicated the signal, the more likely the answer is going to be something like "make sure that when your signal goes away, you are left no further long or short than (some interesting function of the slippage and commission)."

Having said all that, I should just pick a model and go forward, but I don't want to pick a weird definition of "signal", for example, and have a bunch of people jump on me for being an idiot.

01-17-2002, 05:58 PM
You would be OUT if there were no signal either long or short but your expectation of vol is very high.

01-18-2002, 04:11 PM
How about, when it signals a long signal, the next exit/reverse/short signal is likely to come at a time after the long has shown me enough of a profit to cover commission and slippage.

I guess another factor is the frequency of signals. If you are expecting to get one in five minutes, and it will come 50/50 up or down, chances are you should just stay in.

So the real answer is probably you should start taking off the position one contract at a time, at random times, if commissions are bigger than slippage, or something. Meaning, the answer is pretty self-evident - and I had already solved it - but living in a world made up of traders like Dr. Bill, I get nothing but grief for even asking such a question.

And then if I further tell someone "I think a small part of my edge is that sometimes I don't exit," I get laughed out of town. But it seems I am correct, thank you.


01-18-2002, 04:11 PM