CAPITAL REQUIREMENT

Discussions about Money Management and Risk Control.
Post Reply
ES
Roundtable Fellow
Roundtable Fellow
Posts: 97
Joined: Mon May 05, 2003 1:02 am

CAPITAL REQUIREMENT

Post by ES »

Dear colleagues,

I would like some input or opinions as to how much one should fund an account to trade equitties. i will trade the turtle system with a twist of course but I am indeed a trend trader.

Any comments or suggestions would be greatly appreciated.

if I can start with 100k for example than fine, if 1mm is required than so be it.
thks
eric
damian
Roundtable Knight
Roundtable Knight
Posts: 815
Joined: Tue Apr 15, 2003 8:43 pm
Location: dusseldorf

Post by damian »

Hello Eric,

For a variety of reasons, I can not give you a dollar answer. But I can give you a rule to base your calculation upon:

Rule: the aim of the game is to stay in the game. Always have enough money to place the next trade.

Given that rule, consider:

What is the minimum amount you need to place the next trade and not care if it is a loser? $Min

What is the maximum likely drawdown you would expect? %DD

Your minimum start capital is: $minStart = ($Min / (1 - %DD)) + ericsRiskUtilityBuffer

There are other factors to consider, but this is one way to determine a key constraint.

I got this equation terribly wrong when I started out and had to pull the plug on my trading just before the 'famous' Q4 '04 trend. A trend that would have reduced the bulk of mid '03 - mid '04 draw down. My miscalculation was due to much greater %DD than I had accounted for and a much greater impact on my psychology from placing a trade at trading at near $Min.

There is a certain amount of luck in avoiding a huge %DD early in your career (I am trading again and... feeling lucky)
ES
Roundtable Fellow
Roundtable Fellow
Posts: 97
Joined: Mon May 05, 2003 1:02 am

quetion

Post by ES »

Thanks Damian. i actually took a break from trading to concentrate on my litigation practice which thank goodness b has been great.

i want to fund an account with a group of guys and i'm wondering whether we should fund 100k to get matters started or whether is a waste and just plug 1mm or more.

Ed Seykota took an account from several thousand and turned it into whatever. c.f. surely had the same success. i want to replicate the returns aiming for between 50 to 100 percent per year on the money. I'm currently working on tweaking the drawdown aspect and i've been able to lower these values considerably. so my project is indeed to take an account from say 100k and turn it into several millions within a 5-7 year period. I did investment banking for 5 years so raising he capital along with my injection of course is not an issue. let's face it man. what's the purpose without the same challenge that these guys faced years ago.

You see, it's not about the money any more. there's no toy or asset our there that i could by that would make may any happier than the present.
I did that crap nice cars, you name it, charter private planes. i want nop part of that crap anymore. i just want to take an account from x to x(10)
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

I find that the Parameter Stepping feature of Trading Blox, helps a lot with these kinds of inquiries. Once I've got a set of trading rules (a "system") that I like, and a set of parameter values that I like, and a portfolio of tradeable instruments that I like, I run a bunch of Parameter Stepping tests.

I'd go to the Global Parameters tab and click on the "Step" checkbox next to the Starting Account Balance ($) parameter. I'd step it from $10,000 to $2,000,000 in increments of $10,000. That's only 200 runs, it won't take very long. A version of this idea, made smaller for easy display over the web, is shown below. (I happened to simulate a futures system rather than stocks, but the testing procedure is the same and the results show the same general trends).

With lots-of-starting-capital, the system can take every signal. As you reduce the amount of capital, the system can no longer afford to take every signal and so the number of trades drops. So does the overall performance. The question is, what can YOU live with?

The worst case tests using different starting dates, will also provide valuable information. They are covered in one of the tutorial movies on the TBB website.
Attachments
blort.png
blort.png (52.77 KiB) Viewed 13257 times
Jake Carriker
Site Admin
Site Admin
Posts: 1493
Joined: Fri Sep 12, 2003 10:32 am
Location: Austin, Texas

Post by Jake Carriker »

To augment Sluggo's suggestion a bit, here is a twist that I use.

Notice that using exponential betsizing it isn't to long before (given a profitable system), even with a small amount of starting equity, you are trading much larger sums. This is great when we want to make as much money as possible. It is not so great when we want to test for what percentage of all possible trades are taken versus rejected for a given account size.

To solve this problem, I wrote a money manager block (included below) that sizes every trade using a constant account value equal to the starting equity you input in the global settings.

BE CAREFUL - all the metrics that are calculated from a compounding equity curve (MAR, Sharpe, etc.) are now rendered nonsensical. Don't pay attention to them. However, now the linear equity curve makes more sense than that ski-slope-from-hell you see when you use exponential betsizing. In fact, if you do one run using exponential betsizing, and then switch over to the constant equity method and repeat the test, you might note how similar (identical?) the logarithmic plot of equity for the former is to the linear plot of the latter. This is, of course, given that you use a large enough starting equity that the same trades are taken in both tests.

Another note about my money manager block: I give you an option to truncate to the nearest contract (share) or round when sizing positions. Lots of folks think that you should always truncate and that rounding is a bad, bad risky thing to do. I can think of some reasons that might not always be the case. In any event, for my blox, the choice is the user's.

**************

Slight change of subject. You mention that you plan to trade equities. Given the rather fine granularity available in stocks (except for special cases like Berkshire Hathaway and some non-US stocks), appropriate starting equity may be more a function of trading costs than anything else. Then again, if you are nitpicky there may be a few other things.

For example, let's say you want to be able to fine tune your risk to within 1/10th of 1% of your optimum per trade risk target. Let's assume a per share stock price of $50. So, we need to do the math that answers the question, "$50 is .001 of what number?" Easy enough, it is $50,000.

If you want to account for initial drawdown as Damian suggests, you can do that too. Maybe you want to make 10X on your account within a relatively short amount of time. Being a realist, you know you can't do that without some pretty exciting drawdowns. So, you count on a 50% drawdown first crack out of the box. Now you have a minimum starting equity of $100K.

I think you probably get it. Plug in your own assumptions. Maybe you want to trade $400+ stocks (Google anyone?). Maybe you want to count on a 75% drawdown because you have big brass ones and plan on betting to make it count. Maybe you want to hit your betsize risk target to within 1/100th of 1%. Any or all of these factors might dramatically change the amount of starting capital you think is best.

Good luck!

Jake
Attachments
Constant Equity Volatility MM.tbx
Money Manager Blox for TBB
(2.63 KiB) Downloaded 927 times
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

Jake, one thing to consider is the "minimum ticket charge" at some stockbrokerages. Suppose Bartholomew wants to trade a $20,000 account, risking 2% per trade. If his system tells him to buy a $10 stock, the correct position size is 40 shares. If his commission rate is 1 cent per share, that's 40 cents commission. HOWEVER if the broker imposes a "minimum ticket charge" of $5.00, Bartholomew's effective commission rate on this trade just jumped 8X.

Fortunately, several testing platforms can simulate this, so Bartholomew can test it out beforehand and see whether it makes a big difference, or not. The most flexible program in this regard is probably Wealth-Lab, with TBB in second or third place. WL actually lets the user supply a program to calculate commissions according to whatever arbitrary algorithm the broker uses.
Jake Carriker
Site Admin
Site Admin
Posts: 1493
Joined: Fri Sep 12, 2003 10:32 am
Location: Austin, Texas

Post by Jake Carriker »

Sluggo, yep that's right.

I hinted at this kind of thing with my comment...
...appropriate starting equity may be more a function of trading costs than anything else.
However, I didn't expand on it.

Seems like TBB *almost* handles the minimum ticket charge situation via the "Commission per Trade ($)" global setting. However, it isn't quite smart enough right now to provide the capability to compare that field with the "Commission per Stock Share ($)" field multiplied by the number of shares for the trade and then debit the greater as the commission.

Currently, TBB will just add the per trade commission to the per share commission.

I opened up the TBB user guide in order to do a nominal amount of research before posting this reply, and I found something that surprised me. Since I haven't done much stock testing with the application, I hadn't noticed it before. The per share commission is quoted as a round-trip rate. Stockbrokers hardly EVER quote commissions like this. They are almost always one-way.

To be fair, at least the manual does warn you about this non-intuitive behavior.
NOTE: Most stock brokerages charge commission for each one-way trade, once for the entry, once for the exit For Trading Blox, use a Commission per Stock Share of double the per share commission charged by your broker for one-way trades.
I wonder why they choose to do it this way, rather than charging one way commissions? That might require less of a warning about how this function probably works differently than most users suspect it might.

Jake
Tim Arnold
Site Admin
Site Admin
Posts: 9015
Joined: Tue Apr 06, 2004 1:41 pm
Location: Boston, MA
Contact:

Post by Tim Arnold »

We could change the commission to be per share one way.

We could also change the per ticket cost to be a minimum instead of additive.
ES
Roundtable Fellow
Roundtable Fellow
Posts: 97
Joined: Mon May 05, 2003 1:02 am

DD

Post by ES »

Dear folks,

It's always interesting to go back in time and review what one has written. Thus, has anyone tackled the issue of minimizing drawdowns? or perhaps tweaking ATR's for exits or perhaps using some other indicator?

Any comments or suggestions re dradowns would be greatly appreciated

best
eric
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

I am sure that many will disagree. It could well be that I am 100% wrong - I have only been at this game a few short years. But viewed as a stand alone investment, in my view it is probably fallacious to believe that drawdowns, in a long term trend following model, can be minimised in any way other than using a small position size, a significant capital base and a large, broad and very well diversified portfolio. I have not tested LTTF on just stocks - but I would not have thought that just trading stocks helps out much on the diversification front. And given the noise in the markets, I think one probably has to trade very long term to remain outside the noise or very short term. And even then sod’s law will ensure you hit a far stickier patch than you would have predicted.

But very short term probably involves greater skill than I am capable of – I always think of Jim Simons and his giant computers programmes, battling against the computers of the other power houses at this end of the game. I realise that I’m just not up to that particular challenge.

But of course, the lower the bet size, the lower the CAGR.

Others will argue until they are blue in the face, usually on the evidence of backtesting rather than long term achieved results, that it is possible to have both high CAGR and low drawdown. They will have tested multiple systems together over multiple portfolios, they will have come up with individual portfolios on individual systems, they will have done all manner of other things to convince themselves that this enviable outcome is within their grasp.

And sure, may be it is for a few short years. But again and again I come back to thinking about the long term results from the trend followers who have been in this game very long term. None of them have achieved outsized returns over a long period of time without outsized drawdowns and that includes the best of them.

Old European helpfully reported the changes recently made at JWH and Dunn – most of us will have been expecting such announcements given recent performance at those two houses. Now I know that Dunn in particular is noted for not caring about drawdowns – just returns.

But what I am trying to say is that the great majority of trend followers out there have had to make changes over time, their systems have had to evolve to changing circumstances. Look at many of the disclosure documents. Look at David Druz – one system over all these years, and CAGR of around 18%, but he seems to have had to tweak it. He had a 50% drawdown at one stage (back in the early 90’s I seem to recall) and had to permanently reduce his position sizing. Look at Abraham – he had to re-think in the mid 90’s and by my reckoning will be in the middle of another re-think.

I think I recently read something by Ed Seykota (in a post on his web site – kindly pointed out to me by TK http://www.seykota.com/tribe/FAQ/2006_Mar/01/index.htm. About adjusting bet size to suit your tolerance for drawdowns. And I think that sums it up.

Or at least that is what I am reckoning on.
Last edited by AFJ Garner on Mon Apr 10, 2006 6:13 am, edited 1 time in total.
AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2071
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner »

As a postscript, I wish I could say that I DID believe in a way to tame trend following so that over the long term I could achieve both a high return and a low drawdown. Then I would devote much more money to it. But I don't think any investment strategy can guarantee that. May be Sluggo (who has, by my reckoning, traded LTTF for a decade) or Leonardo or others will prove me wrong and tell us what they have achieved. Although even a decade is pretty short term compared to the longest serving trend followers out there.

As ever, I continue to think about my personal trading and investment strategy. And regrettably I believe that it is essential to trade another or other strategies alongside LTTF. I trade IPOs for instance but have spent a few years looking at other hedge fund strategies and hope to adopt one or other at some stage.

If LTTF were to be my only trading strategy or investment, I would be placing the majority of my funds in short term government bonds or money market funds and trading relatively aggressively with the rest. But I would then be trading for higher returns and higher drawdowns, with the security of knowing that whatever the drawdowns I would be unlikely to blow up totally. Even if I were to trade a combination of systems which had shown high return/low drawdown performance in well conceived and applied backtesting, I would still be expecting the 100 year wave to batter me at some stage.

It is interesting to note that some of the CTAs out there take the same view. Some combine trading options on the S&P with trend following for instance. Maybe that is the way to go: trade a few complimentary strategies rather than trying to achieve the impossible with one strategy.

But don’t listen to this cynical, cautious old dog – there are far greater experts out there than I.
Post Reply