Extreme Deviation Automatic Price Targeting - "X_DEV" - Bitcoin

[stiv]

Hi Tom,

As expected, VLNK came back some yesterday. It's bad enough trying to play with the penny's, but trying to catch a parabolic top is even harder. I don't think Myst and Yong have it quite ready for day trades.

I think it is one of the strengths of x-dev, as opposed to (my limited knowledge of) AIM, that it will get you fully in or out of a position in order to maximize total return. It only will hedge the bet so far or so long before committing everything one way or another. You either have to wait for the next action to occur or override the system to take advantage of an extreme trend or significant trend change.

As Myst described, if the system is all cash and a trend change occurs to the upside, one should consider restarting the position at a partial vest.

While x-dev is programmed to take action at the extremes, it typically won't try to handle the ultimate extremes. It will only do so at the expense of total return. It just won't pay. And since the extreme extreme is usually short lived, they usually revert back closer to the mean. Unless of course it is a total collapse.

Now, you can make x-dev go for the ultimate tops and bottoms, but it will sacrifice significant performance to capture them. One way to program x-dev is to open up the bands to hit the highest highs and lowest lows. This usually is not the most profitable set up, since it will ignore the real profit potential territory which is inside those extreme solitary bands.

Another way to try to hit the very tops and bottoms is to adjust the buy and sell multiples. To illustrate with VLNK, if I adjust the sell multiple, I can have x-dev be relatively more or less aggressive when selling a position. The higher the multiple, the more aggressive it will be - the sells happen quicker and at higher share amounts. The lower the multiple, the slower the sells happen and with fewer share amounts.

So with VLNK, if I lower the sell multiple, I can stay in a position longer and try to catch the blow-off top. But as you will see, you end up missing the profit on trades that could have happened more frequently further inside the bands.

To illustrate, here are a few charts that show different sell multiples for VLNK, all else being equal. Of course, every issue is different, but for VLNK, the aggressive seller will get the profits. Note how with a sell multiple of 1, I can get that extra 40% move on the last trade, but I would have to give up nearly everything else from earlier trades to do so.

Sorry for the bandwidth feast...


Sell Multiple = 13



Sell Multiple = 10



Sell Multiple = 5


Sell Multiple = 2


Sell Multiple = 1



Now to clarify, and here's the ironic part of everything I just said, I am not yet trading VLNK mechanically with x-dev. But I have used it exclusively on the buy side for some very key trading decisions to average down my original pre-xd, poorly timed purchase.

Because this is a high risk and relatively low dollar amount penny play, I am looking for a significant trend change to the upside. One that will generate those dreamy pie in the sky returns of a couple of years ago. In essence, I am trying to capture that extra 40% by overriding the system. Exactly what I just argued against doing because you can't beat the x-dev system!!

Didn't I just say, "the aggressive seller will get the profits"? Doh! Someone please hit me on the head with the big xd hammer.

stiv