Buying Stops - inefficient use of capital

Hi all,

I have a general question for people who use buy stops. At present, I trade a system that produces a list of potential stocks to trade. That list of stocks would also include a threshold pricing level for which the trades are to take place only if the stock price rises and then crosses over that threshold level.

I have no problems with setting up the orders in such a way so that the trades are triggered as required. However, the problem for me is that out of 5 (say) potential stocks listed by my system, only 1 or 2 may rise above the threshold level triggering the trade. Because at the time of placing the conditional orders, I have no idea which (if any) will be triggered, I have little choice but to issue the conditional buys on all 5 in order to corner the 1 or 2 stocks.

The problem is that every conditional trade I place will “reserve” a lump of my capital at the time of placement, irrespective of whether the trade actually takes place or not. This ties up my capital and prevents me from trading other opportunities. On some days, it is entirely possible to have no actual trades take place from the 5 conditional trades that I put in, which is very inefficient usage of my capital.

What I would like is for my broker to “reserve” capital only for the trades that take place, and not at the conditional stage. Therefore if I have capital to only take 2 trades then once the 2 trades have gone through, the broker can cancel all remaining conditional orders. That way, I can place as many conditional orders as I like (without tying up my capital) and my capital will be used only when the actual trades take place.

Another alternative would be for me not to enter any conditional orders at all, but to just set alerts for the 5 candidate stocks and when any of them rise above the threshold level, I get a SMS and then run to a screen to activate the order accordingly. Unfortunately, this is not practical for me.

The reason I brought this up is because even a very good system can be rendered ineffective if the capital required to trade it becomes exorbitant. Just curious as to whether anybody else has similar problems and how they have gotten around it.

bingk6,
Probably a few ways around it:

(1) Change brokers to one that does not put aside the capital
(2) Use leverage such as CFD's so your capital input level is less
(3) Amend the system with a filter to be more selective in its trades

I devised my Bang For Buck filter to do exactly this in the late 90's. Back then we didn't have CFD's and minimal margin lending that was cumbersome at best. Too many signals, not enough capital. I required somethiing that would somehow sort some of the signals out to suggest possible better performers and possible secondary performers.

A discussion on the Bang for Buck can be found HERE

I have had a lot of people come to me after doing various trend following courses being bantered around by two well known educators. The issue is that they would get 10 signals every single day from the system. They then run into the problem that you have discussed, but also the added problem of then using a discretionary element to select which signals to actually take. The Bank For Buck could also help there as well.

Nick

As I mentioned earlier on RC, the explanation of that formula is not quite correct, and won't give the right absolute figure if anything other than $10,000 is used. However, relative values will still be correct, so it won't affect which stock would get chosen.

The comment that dividing the result by 100 is to convert to dollars and cents is not correct. All values are already in dollars. It's actually to convert it to a percentage of equity, which only holds if the equity is $10,000. The final figure is independant of both price and equity, just giving a percentage gain.

The formula simplifies down to:

BFD = ATR(200)/C*100;

which gives the percentage gain that would occur by an ATR(200) move in the price.

To get actual dollars per bar for a given equity, use the formula as described in the document but don't divide by 100 at the end. In other words:

BFD = E/C*ATR(200);

E/C is the number of shares, so this gives the total dollar increase from an ATR(200) increase in the share price.

Cheers,
GP

what broker wouldn`t put money aside when placing an order?

Every futures broker on the planet as a start...

Nick,

Many thanks for your feedback. I will look further into selecting a broker that wil not reserve funds at the placement of conditional orders. IB entering the Aust market could just be what the doctor ordered.

Secondly, I am not terribly keen to use CFDs at this stage unless my system can produce a MDD% of less than 15%. From what I understand, both gains and losses are equally magnified with the usage of CFDs. Using a margin of only 20% with a MDD% of 20% will clean me out, irrespective of how much I've made prior to experiencing the 20% MDD. Makes me wonder how people trade their CFDs with their 3-5% margins, as a 5% MDD will clean them out and a system with only 5% MDD is a top class system by any measure. How good do you have to be to be able to trade CFDs successfully ......

Finally, yes, I am already using your BFB to filter the signals and to go for the trades that show more promise.


GP,

Thanks for your reply. I guess the question as to whether to divide the final answer by 100 will depend on whether ATR(200) returns its answer in cents (in which case you'll need to perform the divide) or if it returns its answer in dollars (in which case you don't need to perform the divide). So the answer to the question really depends on which Software package you run the equations in. As you have pointed out, its all relative and the importance is really to ensure that the same equations are applied uniformly so that we can pick out the more volatile stocks.

Nick Radge said:

Every futures broker on the planet as a start...

Click to expand...

I thought he was trading stocks on the ASX, not futures

at least with IB you only put up 50% of the trade and I think your positions are counted towards buying power as well.

bingk6 said:

I guess the question as to whether to divide the final answer by 100 will depend on whether ATR(200) returns its answer in cents (in which case you'll need to perform the divide) or if it returns its answer in dollars (in which case you don't need to perform the divide).

Click to expand...

If the ATR was in cents and you wanted the result in dollars, then yes you'd need to divide by 100.

I would have thought though that ATR would be in the same denomination as all other prices in the system, which I would expect to be dollars rather than cents (for currencies that use dollars and cents of course). But maybe some systems do store prices in cents.

Cheers,
GP