Ralph Vince Interview

www.bettersystemtrader.com

Ralph Vince Interview

Better System Trader Episode 011

Andrew:

Hi Ralph, thank you for joining us today.

Ralph:

Hi Andrew.

Andrew:

Anyone who has done some research into position sizing or money-management would've come across
your work so can you give us a little bit of background into how you started out in the industry?

Ralph:

Well I actually got a job as margin clerk before margin on accounts that trade short option positions was
computerized. It was actually done by hand for a while and that's kind of a baptism by fire. But thats the
short story of how I got in to it and one thing led to another and I have been very lucky and I found myself
working as a programmer for Larry Williams a few years later and he was in the Robbins World Cup 1987
championship.
We also had some managed accounts at the time and there was some disparity in the performance of
those. And at the same time we were trying to implement the Kelly criterion in his World Cup trading and
there were some things I noticed about that that really didn't make it applicable to trading, that I was
trying to solve and the solution to that was also the solution to the disparity we were seeing in some of
the managed accounts.
So that's how I got into the trading business and how I got... kind of turned into the direction of moneymanagement. And the farther along that line I got the more of an obsession it became and here we are.

Andrew:

So I think 1987 World Cup trading championships, was that the year that Larry did 1000+ percent return?

Ralph:

Yeah. I think we had about a 900% drawdown in there too but he took $10,000 to over 1 million and it
was a wild ride but he did it and it was very impressive.

Andrew:

Okay I might come back to that later and ask you about that a bit more but just to start off with you, I
would like to start off with a few basic questions before we move into the more advanced stuff, so how
important is position sizing in a trading model?

Ralph:

Well, I mean to me, I am biased this way because as I said it's an obsession to me but I believe it's the
single biggest factor in anyone's trading, probably the most overlooked. It tends to not be as interesting
as looking at the price charts and price patterns and if you bought here and sold there. But in terms of the
importance on one's bottom line I think it is second to none.

Andrew:

And I suppose there are a number of position sizing options out there. What factors do you need to
consider when choosing one that suits you?

Ralph Vince Interview www.bettersystemtrader.com

Page 1 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Ralph:

Well, I always say there is really There is three of them but they are really directed towards one. And
the single most important thing is criterion. What criterion/criteria are you seeking to satisfy in your
trading? If you are looking to shoot the lights out as Larry did in 1987 that is one criteria. You might have
a different criteria. You might say I want to get the greatest return I can without exceeding a 20%
drawdown. Another guy might have another criteria that says I just want to make sure that when the boss
calls me into his office three months hence I am profitable, that I have got my hurdle plus anything.
So there is different criteria, different traders out there and most people never really identify their criteria.
I mean the criteria for let's say a pension fund manager is going to be different than the criteria generally
for the individual speculator. And regardless of criteria there is two supporting legs to that and the first
one that is unavoidable is horizon. And I can get into that more but the horizon, how many periods you
are seeking to satisfy this criterion over is absolutely essential. Absent that, you're really going to have
trouble satisfying the criteria and I don't like that but that's just the way the math works this one out.
And the second is let's say to what degree of confidence do you want your solution to this criterion to be?
So that's really the basis behind anything in money-management I think. In detailing the criterion and
along with that, over how many holding periods, over what time periods are you looking to satisfy the
criterion over?

Andrew:

Right. So most people would know you for your work on Optimalf. So can you explain a little bit about
that concept and how it relates to position sizing?

Ralph:

Sure. Optimalf is a fraction of Let's see, to expose to risk such that one would maximize their expected
geometric growth. And that too is a function of horizon and I will explain shortly. When I say it's the point
that maximizes growth, it implies that there is a curve and it would be the high point of the curve. And
the important thing about that is there is a curve with certain points on it of geometric importance to the
trader depending on the criterion and the peak of course is one of the most essential or the most essential
point in defining the shape of that curve.
When I say that horizon is important in terms of Optimalf, in terms of maximizing expected growth, let's
go back to that 2 to 1 coin toss game we were talking about earlier, the .5 probability. We could say that,
Oh, the Optimalf on this is .25." The expected growth maximizing fraction to risk on this is 25%. That's
only true as the number of periods or the number of plays approaches infinity. If we were to play this for
one play I say, Andrew, I'm going to toss a coin and if it comes up heads you win two dollars. If it comes
up tails you're going to pay me a dollar. What fraction of your stake do you risk to maximize what you
would expect to make? The answer is 100%. If we are going to quit at two plays now it's a lot less than
100% and if we were going to quit on an infinite number of plays, it converges asymptotically to the peak
of this curve at .25.

Andrew:

So when you said the peak in the curve is the most important point, are you talking about Is it the most
important point for optimal growth?

Ralph:

For expected geometric growth, yes.

Ralph Vince Interview www.bettersystemtrader.com

Page 2 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Andrew:

Right. So what if your criterion is perhaps to manage your drawdowns? I know we have some prop traders
who listen to the podcast and I guess their focus and for other people that manage money for other
people, their focus is probably more on keeping the drawdowns lower rather than getting a huge return
so how does this Optimalf concept adjust to a different criterion?

Ralph:

Okay well let's say, drawdown is a different animal altogether and it too is horizon specific. But if I may,
rather than addressing drawdown, let's say, I want to trade such that I have a floor on the account value
and it's 80% okay of where we are today. So I am not addressing drawdown per se but I am in fact
practicing let's say a degree of portfolio insurance. And we are trading one component and it is our 2 to 1
coin toss as it would be okay. Again, this is just a simple analog for a discrete distribution of outcomes
from trading. And in this case it is just one component but for simplicity. And we say you know I don't
know what in the hell my horizon is. I plan to live a long time and just keep on trading and trading and
trading and trading so my horizon is going to the asymptotic value .25. Okay that's fine.
When we start out there were to go on a risk .25 but you're looking to say: you know but I really can't
concede dropping below 80%. Well now you can't trade .25. You can trade .2 and if you lose in the first
play you are down 80%. And what we have is As one loses money, one covers a path from the peak or
some point less than the peak towards zero as the equity in the account diminishes. And as the equity in
the account increases, that path increases, the amount you are committing, the fraction you are
committing increases towards the optimal point. So we are no longer on just the single point on the curve
but we are actually traversing a path on the curve. And that's very important in understanding this because
some criteria, let's say expected growth maximization requires us just being in a stationary point. Other
criteria requires traversing a path through this surface.
And as I say the surface itself or in this case the surface is one-dimensional because it is only one
component, can be solved through some very complicated mathematical and computer processes or we
can estimate it very simply.

Andrew:

So talking about this curve, how do you or what part of the system actually dictates the shape of that
curve?

Ralph:

Well actually the distribution of outcomes dictates that curve and so from one component we have a two
dimensional curve like a curve you drawn on a piece of paper. And it is bound on the left at zero and on
the right at one and representing the fraction of our stake we would risk. And I translate that into a
number of units whether they are shares or futures or whatever by dividing the largest loss of the system
by that fraction.
So if my largest loss is lets say you are trading the system and the largest loss is $1750 okay and my
Optimalf is .27 so I have a number of 6481. So for 6481 in account equity I'm going to trade one unit, one
contract or 100 shares or whatever to go from a fraction to a number of units that represent that fraction.
If we trade two components, we have two axes that are aligned at 90 to each other, both bound at 0,0
and go around clockwise I would be at 0,1,1, 1,0 and back to 0, 00. And within that field, within that
manifold I have a curve in 3-D space and that curve would have a peak.

Ralph Vince Interview www.bettersystemtrader.com

Page 3 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Let's say I am trading two coin tosses that paid 2 to 1. Now my peak in that curve is .23, .23. And so if I
trade N games and have a in N+1 space or if I have a surface in N+1 space that has a unique peak to it.
And another rule of thumb regarding let's say the peaks is this; in the absence of negatively correlated
instruments which are virtually impossible to find and put us into a different universe, we find that the
sum of the F values as the number of components increase, that peaks out that 1. So if I have one coin
that pays 2 to 1 and I have a curve in 2D space, the peak is at .25. If I have two of these games
simultaneously, I have a peak at .23, .23 or the sum of that is .46. So I'm going to both come up losers
here, I am down to 46% on the first simultaneous play.
Now as I keep adding more components or games to it I find that the sum of the up values peaks out at
one. Which is quite interesting because it says that the Optimalf for a very large portfolio of stocks is a full
100% commitment. But it also helps in terms of rules of thumb on these things, along with the earlier rule
we discussed of taking our ability of a winning outcome and dividing it by 2 as I adds components to that,
that diminishes for the individual component because I think the sum of all of those components peaks
out at 1.
Andrew:

All right okay, I think you may have touched on this a little already but I would just ask for clarification. So
you said the curve or the surface is created by the distribution of returns so should it be recalculated after
every trade?

Ralph:

Not really. I mean unless it grossly alters the distribution of returns. Ideally, I mean in an ideal world, you
estimate your distribution of returns and as the returns come in they fall within what you expect given
that distribution. We have events like 2008 and everybody's distribution goes out the window which is
also why let's say we build a distribution off of past empirical data, I personally choose to amend that to
be not quite as benevolent as the past.

Andrew:

All right, okay when I guess that's probably one of the main criticisms with optimalF is that it's based on
historical data and you never really know when the biggest loss is going to hit you so how do you account
for that?

Ralph:

Well, two things first of all part of what I have been pursuing more robust methods and let's say this
profitability, the sum of the abilities of the winning outcomes divided by 2 is the best guess estimate for
where things will be in the future because now it's totally dependent on what the probability of winning
trades is or holding periods is in the future which can be fairly accurately modelled okay.
Let's say in a trend following system might win 30% of the time over any considerably long time period. If
my future is a considerably long time period, I would expect 30% of those trades to be winners and if it's
the only system I am trading and my criterion is to maximize expected geometric growth then I will take
that .3, divide it by 2 and say, Oh, .15 is my best guess for the fraction I should risk on this to be expected
growth optimal for an infinite number of plays on this."
Now the largest loss comes in terms of translating that fraction into a number of shares and contract. And
I myself find that it pays to use other instruments to rather cement that in whether it be options or
anything along those lines to cement in what the largest loss is. And tail risk is something everybody is

Ralph Vince Interview www.bettersystemtrader.com

Page 4 of 18

Ralph Vince Interview

www.bettersystemtrader.com
concerned with in dealing with these days, there is various ways to address that but it is critical. If you get
a new largest loss you are going to lose your footing and lose your balance and I think that's the truth of
anything.
Andrew:

Right so that was interesting that you said that if there is something that adjusts the distribution of
returns, that that's probably when it should be recalculated.

Ralph:

I think so. I mean 2008 certainly everyone had to go back and rethink everything after that.

Andrew:

Yeah. You have already touched upon the nature of the curve but I would just like to ask a little bit more
on that. I've got a quote here that you made in the IFTA journal, if I could just read that to you. It says,
The optimal F calculation provides a bounded context of studying the nature of the curve whose optimal
point iethe peak represents the correct fraction of us take the risk to result in the greatest geometric
growth asymptotically. There is the nature of the curve whose bounding allows us to study the different
phenomena of the curve as well as provide a context from which we pursue criteria other than mere
growth maximization."

Ralph:

I never said that, I never said that. [laughs]

Andrew:

Well you didn't say that?! [laughing] A different Ralph Vince was is it?

Ralph:

Yeah, I didn't say that! [laughs]

Andrew:

So you have already mentioned a little bit about the shape of the curve, I just wonder if you could tell us
a little bit more about the practical applications of that.

Ralph:

Sure. I mean again I get back to criterion as dictating what we should do on that curve. The peak that
optimal F point as I call it or you can call it the Kelly criterion point which is a non-scaled optimal F point
is a critical point but there is other critical points on there. One of them of course is the inflection point
and the curve tends as we increase the fraction away from zero, it is accelerating to the upside and at
some point before it reaches the peak that acceleration diminishes and we go from let's say something
that is concave up to concave down. There is a point of inflection that is less than the peak which would
represent where marginal increasing gains diminish with respect to marginal increasing risk and I think
that's a critical point for money managers.

Ralph Vince Interview www.bettersystemtrader.com

Page 5 of 18

Ralph Vince Interview

www.bettersystemtrader.com

Image from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2364092 with permission by Ralph Vince

Another critical point that is shy of the peak, and may actually be the optimal point for most traders
aggressively would be that point that maximizes gain with respect to drawdown. And we know that
drawdown is linearly correlated to f value, that is the greater the f value, the greater the drawdown. So
let's use the f value as a proxy for drawdown and thats probably a pretty accurate one because as soon
as the largest loss is hit we are going to see at least an F percent to drawdown. And if we draw a straight
line from 0,0 that would be tangent to the curve, that point of tangency is somewhere shy of the peak but
that point of tangency represents greatest growth with respect to drawdown, greatest growth with
respect to risk.

Image from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2364092 with permission by Ralph Vince

And many traders would say, Well this is your optimal point," not exactly the peak but this tangent point
here." And a fourth very critical point is a point right of the peak where the curve drops below 1.0. The

Ralph Vince Interview www.bettersystemtrader.com

Page 6 of 18

Ralph Vince Interview

www.bettersystemtrader.com
altitude of the curve represents what we make on our stake as a multiple. And if that multiple drops below
one then what we make on our stake over those periods is a number less than one.
And here's where it gets really interesting into all kinds of other applications, just as we have an F value
between zero and one, we can also say that I don't want to get into that geometry, that's really That
is going to go off on a bad tangent here Andrew. It is going to be very time-consuming and really not
germane to what a traders interest would be in it. Let's leave that is going off in a bad direction there.
Andrew:

Well maybe we can move into multiple strategies in the portfolio because many traders look to reduce
their risk by diversifying with other strategies or two different markets. So what impact does this type of
diversification have on risk?

Ralph:

Oh, it's a great idea! I mean if I can trade multiple strategies and multiple components, oftentimes I will
get a better diversification with just multiple components. Let's say I am trading just a trend following
system, let's say it is some type of channel breakout system. I can trade that on let's say five different
markets or let's say I trade five different systems on one market and those five different systems are quite
disparate indeed, they are not all trend following systems. It's a lot easier to find low correlations on the
same component between different systems than it is to find low correlations on the same system
between different components. As a general rule, I mean it's not always the case but at any time you can
add more viable systems you are better off.

Andrew:

Right. So what if you do have multiple systems, does the optimal position size for a single strategy change
when you are including these other strategies in a portfolio?

Ralph:

Absolutely. The shape of the curve remains the same but the peak is going to be at a different location
and hence those other two points we mentioned, the inflection point and the growth risk maximizing
point, those will change as well because those tend to orbit the peak. So it's going to change as well. But
again, there is some pretty futuristic rules for estimating the shape of this surface given the probability of
winning outcomes for these various markets and systems. And we would say, Okay the shape is going to
change but it is readily calculable."

Ralph Vince Interview www.bettersystemtrader.com

Page 7 of 18

Ralph Vince Interview

www.bettersystemtrader.com

Image from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2364092 with permission by Ralph Vince

Andrew:

So from a money-management point of view though, should multiple strategies be managed separately
or are you saying that they should be managed as an entire portfolio?

Ralph:

I put them in I aggregate them together. And if you separate them, there is a little bit of an efficiency
loss. You can do that although you do lose a little bit of efficiency by doing that, you are a little better off
to aggregate them together.

Andrew:

Okay thanks Ralph. I would just like to move on to the listener questions. We've got quite a few to cover
but first I just want to ask you a general question. So your first book came out I think in 1990. Have your
ideas or how you apply them changed over the past 25 years?

Ralph:

Oh absolutely! At first I was solely interested in the peak and solely interested in the asymptotic effect of
that. And I have since come to realize that the whole story is in the curve, the shape of the surface and
how we move about or position ourselves on that surface to satisfy whatever criterion we are looking to
satisfy. It is a much broader application than just saying I want to go and make the most on this money
and willing to lose it all and drawdown be damned. So I mean that's clearly a big change from back in the
late 80s when I was putting that together originally.

Andrew:

So what caused that change in your understanding?

Ralph Vince Interview www.bettersystemtrader.com

Page 8 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Ralph:

Oh, age, age! [Laughs] everything just slows down a little! [Laughs]. I mean the more I looked at it, the
broader it became and the more critical again the idea of horizon, the criteria became the more I looked
and said, Okay, this is applicable." Every single trader out there could use this to his benefit, his biggest
impediment to doing so is being able to articulate his criteria and which also means articulating what his
horizon is. And if he can get past that you've got half of this licked, the calculation aspect is not hard. The
theory of how to apply that calculation aspect is not hard. The biggest obstacle is determining, really what
is your criterion? What is it you seek to gain? For the individual that is particularly difficult. Like I said for
a pension fund it is very straightforward. For an individual it is very difficult for an individual to say, My
criterion is this." It is not easy.

Andrew:

Yeah, that's a great tip to have a good understanding of your criterion first before you get through all of
this stuff, really helps with the outcome.

Ralph:

Yeah!

Andrew:

All right, well thanks for that Ralph. We will just move on to some listener questions. The first one is from
Patrick, I would just read it to you. It says, Some strategy developers have argued that after testing
myriads of position sizing algorithms, one of the surprisingly most efficient ways to protect from a system
breaking down is to stop trading when the equity curve is below its 200 day moving average. What are
your thoughts?

Ralph:

I've tested those kinds of ideas out quite a bit and I found it to be a case of 6 in one 1/2 dozen on the other
so they say. I mean it helps in other situations, in some situations it hurts. Again I look at things in terms
of a path in this manifold, in this leverage space manifold on that surface.
So if you're saying, I'm going to stop trading at such and such a point" okay you are actually traversing a
path on there where at some point in the path it just ends. It can work but there are times when this idea
won't work and you would pay the price for it. But again if we look at just past data we could curve fit
anything especially to the 2008 episode and say, Oh in a 30% drop we shut down trading altogether."
And there is a lot of stuff out there to that effect but not knowing that beforehand I think you could hurt
yourself with it as well.

Andrew:

All right, okay. And there was a second point to Patrick's question actually. It says, If we assume a strategy
will lose its edge at some point, how do you identify that point and protect yourself?

Ralph:

Well you have to be monitoring the outcomes of trends to whether they conform to this given distribution
or not. I have done a lot of work along those lines but you really need some of the stuff of Abraham Wald
who had a book on I think it was in 1944 called Sequential Analysis. And I'm still working on that, I don't
have the answer to that, that's just the direction that I am looking at in terms of solving that is at what
point do the trades no longer comport to the expected distribution of outcomes?
And I think that's very similar to I would say Wald's book that tries to address from a manufacturing
perspective. We would say we make widgets, they come down an assembly line, the 11th which it is
defective, the 17th widget is defective. At what point do we say there is a problem with our manufacturing

Ralph Vince Interview www.bettersystemtrader.com

Page 9 of 18

Ralph Vince Interview

www.bettersystemtrader.com
process here? That things aren't coming along as expected? How do we identify that point as early as
possible with some type of mathematical justification? And again, I would say I'm going to study some of
Wald's stuff for that.
I don't have that solved yet. I am working it so if Patrick gets to it before I do I certainly hopes he shares it
with me.
Andrew:

Okay great. Thanks. So the next question is from Shawn. He says, If you have read Howard bandy's book;
quantitative technical analysis what is your opinion on the safe-f position sizing described therein?"

Ralph:

I have to confess I haven't read it. I should have, I just haven't read it yet and it is the first I've heard of it.

Andrew:

All right, we will move on to the next question which is from Vikas. His question is, "What is some proven
optimal money-management strategies to take a small account to a large account?" I think this one ties
into Larry Williams story as well.

Ralph:

Right, exactly. I mean if you want to do that I would say to trade at the optimal level and he would say
that if you're looking to trade one component you would quite simply say; what's my probability of
winning trades divided by 2, take that result and divide your largest loss by that gives you a number to
trade how much money to put behind each contract or number of shares your trading. But again if you're
going to do it that level you better be ready to be wiped out and at the very least, for some very wild
swings.

Andrew:

Right, yeah.

Ralph:

It is, it is a very wild ride, you are not going to get there on a straight line.

Andrew:

So is that how Larry Williams increased his 10K to 1 million+?

Ralph:

Yes.

Andrew:

In 1987?

Ralph:

Yes.

Andrew:

I think you said he had a 900% drawdown.

Ralph:

Yeah, it is a wild ride, it is a very wild ride! And here's the thing, the better the system the higher the F
value and the higher that F value is, when the largest losses hit you're going to see at least an F percent
drawdown. So the better the system in this type of approach in trying to satisfy this criterion the more
severe you can expect your drawdown to be. It's kind of paradoxical in that sense but it's the way it works
out.

Andrew:

I guess there is no such thing as a free lunch right?

Ralph:

No, no, no, no! I don't think you want to have lunch with that strategy at all! [Laughs] I would just keep
the bag handy and forget the lunch! [Laughs]

Ralph Vince Interview www.bettersystemtrader.com

Page 10 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Andrew:

[laughter] You won't feel like eating when you see the drawdown?

Ralph:

No, no I get one of those two bags and the front seat, the seat right in front of you there and just hang on
to that.

Andrew:

All right. The next question is from Ola.

Ralph:

Okay.

Andrew:

Ola asks, I would like to hear his views on dynamic position sizing with respect to a strategies
performance. If you measure the performance of a strategy by the performance of recent trades, can you
improve the performance and or limit the risk-taking by adjusting the position sizing accordingly?"

Ralph:

Well okay when you speak of let's say improving performance and I say, Okay, what does that mean to
Ola? What does she mean by "improving performance?" Again it's a function of criterion.
Let's say Ola's criterion is to maximize the probability that her equity curve will be at a new high equity
and so all it does is she does things dynamically and as the system loses money, she tends to commit more
money to each subsequent trade up to the optimalf value. So let's say she starts out that a very modest
position sizing. Let's say her optimal up is .25 but she starts out at .03, to minimize her drawdown, she has
a couple of losing periods, a couple of losing trades; that .03 is now .1 and things are still going sour and
so forth.
Eventually she is let's say all the way out .25, the optimal amount which if the system does anything like
it did in the past you will quickly recoup the accumulated loss there. So it's a case of okay, her payoff by
the way, the exact opposite direction of the insured portfolios path we just mentioned earlier in this
discussion where we start at the point and move towards zero as losses accrue, here she is starting near
zero and moving towards the peak as losses accrue. This type of dynamic strategy would help satisfy her
criteria which I am hypothetically stating here as maximizing the probability of being at high equity at
some arbitrary future point.

Andrew:

All right okay. Thanks so the next question is from Brian. He says, Ralph, I admire your work. My question
is about martingale strategies. Probability theory suggests that using martingale methods in money
management is a gamblers fallacy. However, I heard that some very successful hedge funds apply
martingale strategies to double down against losing positions. Have you ever or do you use any martingale
strategies and do they make sense?

Ralph:

Yeah, they do make sense. They have a place in things but again, this is in many ways like the 200 day
moving average on the equity curve. There are times where it could really burn you as well. If you were
trading let's say a martingale strategy throughout the 1990s, you would have far outperformed what fully
invested portfolio might have seen. So it's the kind of thing that in the right market environment, in a
congestive market environment you would be much better off because you are buying more at lower
prices. But in a runaway market, you're going to pay a pretty steep price for that.

Ralph Vince Interview www.bettersystemtrader.com

Page 11 of 18

Ralph Vince Interview

www.bettersystemtrader.com
Andrew:

All right okay thanks. The next question is from James. His question is which of Ralph's book would he
recommend for someone whose eyes immediately glaze over when mathematics is discussed?

Ralph:

I think the original one, portfolio management formulas has the least math in it and covers the concept
pretty basically. Unfortunately I think they all have a good degree of math. The latest one, The Risk
Opportunity Analysis I try to write in such a way that one didn't have to delve into the math although the
math is presented and that is pretty thorough. So he's probably better off with that but really with an eye
towards disregarding the math because again math can either be run by computer or calculated with
heuristics you could perform quickly in your head.
So to James I would say don't let yourself get hung up on the math of it but do cover it for the principles
that we discussed for the inflection point and so forth.

Andrew:

Right. Okay and we've got another question from James actually. He says, I think Ralph had a role in Larry
Williams position sizing model as released in his Long-Term Secrets To Short-Term Trading book 15 years
ago where everything, position sizing revolved around the largest historical loss. Given the markets are
now virtually open 24 hours a day and that the weekend is the only time a nasty gap could strike, does
Ralph think that one can dispense with that safety valve and be a little bit aggressive with position sizing?"

Ralph:

Well first of all, just to back up real quickly, I wouldn't say that weekend is the only time a nasty gap can
strike. You could have a nasty Strike right in the middle of a trading day. The mechanics of the marketplace
is now such that this could happen. And because we haven't seen it doesn't mean we couldn't have let's
say an instantaneous print of -1000 DOW points here in the middle of the day. We don't have a specialist
system that works anymore. We have an ersatz specialist system called High Frequency Trading. There is
also some very, very nasty feedback mechanisms as a result of leveraged structured products nowadays
that's at work in the shadows here. And I am as afraid on a minute to minute basis of a nasty gap as I am
over the weekend I would say no, we certainly can't to dispense of largest losing trade. And think of it
as worst-case outcome. What's the worst case that could happen?
If you are just using largest losing trade historically you may not be enough okay. But largest or worst-case
outcome I would certainly know. Let's say I am long some volatility points okay. I know my worst-case
outcome is going to be the amount of, the premium of those points. And so I could truly truncate my
worst-case outcome but I absolutely can't dispense with it, I mean especially now.

Andrew:

All right okay thanks. Now the question from Ryan which I think you've already covered, what fraction of
optimal if do you recommend is advisable?

Ralph:

Oh, that's a function of the system itself okay, the distribution of outcomes. What other systems are being
traded with it because let's say I trade 1000 components, 1000 different marketing systems, I know that
the sum of the F values is going to total up towards one in 100%. But that assumes that my criterion is just
straight up maximizing expected geometric growth. And if that's the point then the optimal F value you
adjust it could be more than 25% and as I say, what's important when you are trading more than one
component, more than one market system isn't individual F values but their summation because that is

Ralph Vince Interview www.bettersystemtrader.com

Page 12 of 18

Ralph Vince Interview

www.bettersystemtrader.com
the percentage of drawdown you will see if worst-case is manifest along all those components at once
which is always a possibility.
Andrew:

Now a question from David, he asks, Please ask Ralph to explain in non-technical language how an
intelligent advisor or an investor could put your ideas or one or two of your ideas into practice."

Ralph:

Okay well let's say I am trading N components okay. I have N different markets and systems. Now for each
of these N markets and systems I am going to figure out what the profitability of a winning month is or
winning whenever my holding period is for each one of these and I'm going to divide that by 2 okay.
Now I want to take each of those results and I'm going to divide them by N and I will explain why I divide
them by N afterwards. This tells me now what the peak of the landscape is in this N+1 dimensional
landscape. And from that peak I can craft what I would want for my criterion.
Now why did I divide them all by N? Okay if I say, you know, I want to be ready for worst-case scenario
where all the correlations go to 1 and everything goes crazy. Because the more a given market makes a
big move, the more other markets at the same time will make a big move. And the more the big moves
tend to move together. So the more a given market makes a big move, the more correlations tend to
migrate towards one. If I have correlations If the correlation of everything is one then I don't want to be
summing my up values.
Let me explain it this way. Let's go back to that 2 to 1 coin toss game we mentioned earlier Andrew. If I
am doing one of these, just one of these at the time my asymptomatic peak, my F value is .25. If I do two
of these very same games simultaneously my peak is at .23, .23 or a sum of .46. That assumes that those
two games are totally independent. And if I keep adding independent games, some of those F values for
all those games simultaneously will turn towards 1.
However, in the case of two coins, they are independent and my peak is at .23, .23, .46 total aggregate
exposure, if those coins let's say have to be perfectly correlated and I don't know this, then the actual
peak is at .125,.125 or the same as playing one game at .25. But I am at .26. I am at way beyond the peak
okay. The last .21 of exposure I have taken on, I am actually paying a price for it of having less return,
greater drawdown. I am really hanging out on this. So if the correlations all go to 1 okay I want to trade
however many coins I am trading right at .25 okay.
Similarly this is why I am dividing all of these components; first I am dividing the profitability of winning
holding period by 2 and then I'm taking that result and dividing them by the number of components. And
this gives me my estimated peak in this landscape should correlations all go to 1.
Now, depending on what his criteria is okay will determine what we do with this information. If his criteria
is, You know, I really want to maximize my probability of being profitable in an arbitrary point in time."
He's going to migrate from an area close to 000 towards that peak. If his criteria is to truncate drawdown
point, he's going to migrate from the peak towards 000 as losses accumulate.

Ralph Vince Interview www.bettersystemtrader.com

Page 13 of 18

Ralph Vince Interview

www.bettersystemtrader.com
So from the simple calculation he can determine where the peak is in a very conservative sense, in a sense
of this is the peak when all hell breaks loose and craft a framework, craft a path through that to satisfy
whatever his criterion is. I hope that is a simple enough explanation.
Andrew:

Yeah, that's a great answer I think Ralph. So let's finish up with one last reader question. This one is from
Vikas as well. He wants to know your thoughts on betting more aggressively with the profits of the year.
So he says, Let's say I am up 20% for the year. Now instead of betting say 1% of the 120% which is the
new account balance, what if I bet 1% of the original capital and then 20% of the profits basically an
aggressive idea to grow capital? So what do you think about that type of approach?

Ralph:

Well again, it depends on what his criterion is. If his His criterion may be let's say he is up 20% on the
year, his criterion might be if he is let's say trading on a prop desk and he can show us a 20% gain this
year, he is King of the floor, he won't want to do this.
If he is an individual trader, this might sit very well with him. There is a big psychological difference
between trading with money that has been accrued through recent gains versus money that youve put
on the table. And it's a lot easier to be more aggressive with accrued profits than with the money you are
playing defensive with if you put it on the table. So it is a very viable idea for a guy who is looking to trade
aggressively if it doesn't violate what his criteria calls for.

Andrew:

Yeah, he originally asked the question about money management strategies to take a small account to a
large account so I assume that means his criteria is growth.

Ralph:

Yeah, yeah. So similarly that could well work.

Andrew:

Yeah, sure okay Ralph. So I would just like to wrap up with a few quick closing questions. The first one,
what is the biggest lesson you have learned through trading?

Ralph:

Oh, things are not what they appear. This has been a learning process for me from the get go. And I think
even in this discussion there are some questions here that pertain to things I'm still working on, still trying
to solve for myself. And certain things that I have grown to be quite certain of, that idea of the necessary
criterion.
Along with that is also the idea that you really can't accept things because they are mathematically
accepted. Remember we spoke of how criterion is everything in this and that's a function of horizon. And
we said when even drawdown, calculation for drawdown and ruin is a function of horizon. And behind all
of this is the idea of expectation, what you expect to make which is a function of horizon. And let me give
you a quick example.
Let's go back to our coin toss example, our 2 to 1 coin toss example and I'm going to throw in a third
possible outcome that the coin lands exactly on its side okay. If that happens Andrew, you're going to pay
me $1 million okay. So you could win two dollars if it's heads, lose one dollar if it's tails or lose $1 million
if it lands on its side. Okay, do you want to play the game? And if so how many times do you want to play
it?

Ralph Vince Interview www.bettersystemtrader.com

Page 14 of 18

Ralph Vince Interview

www.bettersystemtrader.com
And you might say, Oh, winning $2, I would play it all day long. That coin is not going to land on its side."
Okay and that's probably very true. If I play 1 million times it may land on its side one time though and
classical expectation tells us, no, this is a losing game don't play it. And Andrew says, What the hell, I'm
playing it three times. It is not going to land on its side. I am going to play it three times and I'm going to
make some money on this."
So the idea then becomes, okay for a given horizon is the far outliers, do we even expect those to come
into play and the answer is no. We don't expect them to come into play because there is We don't have
the principle of ergodicity at work. As individual traders trading over let's say Q future periods, there is no
such thing as ergodicity here where if we say, Okay 1 million guys are going to come into this coin toss
thing and each guy is only going to play three times, some guys are going to get zinged and it is going to
land on its side okay. But for you Andrew, you're not concerned about those other guys. You are
concerned about Andrew over three coin tosses so you are really not concerned about that then coming
up on its side.
And this idea by the way, I contend is what belies much of what human beings do contrary to what classical
expectation tells us human beings should do. It is the reason why you would get in your car and you will
drive to the office today even though you could get killed doing it. You don't expect to get killed driving
to the office today okay. You really expect that if Im going to do it 10 million times, I will get killed doing
it but I would end up doing it 10 million times.
And so this idea of stopping times in mathematics and classical expectation really do not pertain to us as
individual traders. And this is really very important and it too is a function of criterion. So right about here,
so let's just go back to the coin toss and say you know Andrew, you are going to play this thing 100 times.
You are convinced it is not going to land on its side okay because 100 times is way too small of a sample
space to worry about it coming up and landing on its side. So what are you going to do to maximize what
you expect to make?
And what you would do then is say, Well, I'm going to completely disregard that outlier that one in 1
million chance of this coin landing on its side and I am going to figure, Hey, it is going to come up heads
with the profitability of .5 and tails with the profitability of .5 and that's it, that coming up on its side is
just not in my field of view here. And you then change the distribution of outcomes that you're using as
input to the calculation of this landscape.
Andrew:

That's a good point you bring up about mathematical expectations. Thanks for raising that one. Second
question, what is the hardest part of trading?

Ralph:

Oh, there is really two parts to it. I mean for me. The first part, and these are both psychological. And the
first part is taking heat on a sizable position. And that's never easy, that's never fun and you have to
console yourself that okay, I have been through this in the past, I have been through 2008, I have this
horror story and I survived that and that's never fun. And I try to have positions going on for long positions,
short positions. My goal is if the day is moving big in one direction I should be taking profit on something
that I sewed at some earlier point. Which also means I am taking heat on another position and that's
always difficult.

Ralph Vince Interview www.bettersystemtrader.com

Page 15 of 18

Ralph Vince Interview

www.bettersystemtrader.com
The other part is along with this there is the requirement of patience. A lot of times these markets just
meander like a lifeboat on a dead ocean going nowhere especially summer months are notorious for that.
I'm not saying that we are in that right now but you know there are times when things just meander and
go nowhere and again you have to look at what happened in the past and say, Okay, this is not a
permanent condition here.
So I try to keep available, charts of what has happened in the past when I have taken heat on things and
what things have meandered and gone nowhere and when I really found myself in a tough psychological
situation I just go look at those charts, just remind yourself, Okay you have been through this before, this
is what happened, this is how this one turned out, this is how that one turned out. Here is the one where
you almost froze to death in Boston waiting to catch the T to meet the margin call but you lived." [Laughs]
Andrew:

[laughter]

Ralph:

You look like Scott of the Antarctic out there but you made it! [laughter] It was a lot of fun!

Andrew:

[laughs] is that why you moved to Florida? To get away from the cold weather?

Ralph:

That Boston experienced just about did me.

Andrew:

Okay great. Sorry were you finished?

Ralph:

No, no, no. There is a funny story along with that but that will wait for another time.

Andrew:

Well, you are welcome to share it if you like.

Ralph:

No, we kind of touched upon it there already. It had to do with cold weather and a margin call in Boston
and that was that. [Laughs]

Andrew:

[laughs] okay. You survived.

Ralph:

I survived it, yeah! [Laughs]

Andrew:

One final trading question. What's the best trading advice you have ever received?

Ralph:

You know, you got to keep coming back. You really can't let this thing defeat you. At some point everybody
finds their groove with it if they stick around long enough but it takes a while. It took me a long time. Now
I was really lucky. I had exposure to some great traders, to some great traders and great systems and great
ideas and even at that it took me a long time to find my groove with this, to find out what works for Ralph
okay in terms of to satisfy my criterion, in terms of trading style that fits me. And I think I should include
that everybody's different okay.
There are certain truths out there, I mean one of the truths is you are somewhere in this manifold as we
discussed and you are in there at the point and probably moving around and paying the consequences
and reaping the benefits of that inadvertently so why not use that to your benefit okay? I mean there is
certain truths to things but what your criterion is and how you satisfy that isn't necessarily the same as

Ralph Vince Interview www.bettersystemtrader.com

Page 16 of 18

Ralph Vince Interview

www.bettersystemtrader.com
mine and everybody has to find their own groove with this and that just takes time and if it doesn't then
you are a very fortunate man.
Andrew:

So you mentioned that you had exposure to a number of trading experts. Who would you think had the
greatest impact on your trading experience?

Ralph:

Oh without a doubt Larry did. Larry Williams. And he knew everybody so there was a lot of cross
pollination going on there but this is a guy who didn't sleep okay. This was a guy, he didn't sleep! He did
not leave a stone unturned. Had I not had the good fortune of running into him Andrew I think I would be
punching the time clock at the Ford engine plant here right about now starting the night shift. I was really
very fortunate to have run into that guy, he was a free thinker and part of his freethinking spawned a
thread, an avenue that I went down in my life without realizing I was going down it that opened up all
kinds of doors and had I not done that I wouldn't have encountered you and wouldnt be having this
discussion. So I mean I am really very grateful to have met the guy you know and to have had his influence
on my life which has been a lot more profound than even just the trading ideas which were profound
enough.

Andrew:

It's funny how things happen like that. They can take you down a different path altogether.

Ralph:

And you don't see it when it's upon you!

Andrew:

Do you wake up and say, Holy crap, how did I get here?"

Ralph:

Oh yeah! Every day! Every day when I see that cinderblock on my chest of those positions that keep adding
up against me another cinderblock I go, How the hell did I get here?" [Laughs]

Andrew:

Do you think of Larry? Do you blame him for that?

Ralph:

[laughs] I do, I blame him every day! [Laughs] I would have been so much happier if I was just punching
the clock right now, wow! [Laughs]

Andrew:

[laughs] Well we won't mention that to Larry.

Ralph:

[laughs]

Andrew:

So do you have a favorite trading book?

Ralph:

I have a number of them and they are all pretty old. But I do have a lot of them and some that I have read
recently that I really, really enjoy. I think one of the ones I really enjoy it a lot was Prector's book on Elliott
wave just because of the way it got me thinking about things in a lot of ways. I like the old Welles Wilder
book because it was so mechanical and at the time the personal computer was just coming on, here was
some very mechanical approaches and so many things had spawned off of that so it is just some of my
favorite books out there.

Ralph Vince Interview www.bettersystemtrader.com

Page 17 of 18

Ralph Vince Interview

www.bettersystemtrader.com
And of course the two, the real cornerstones of technical analysis to me are the chartgraph point and
figure book and of course the bible Edwards and Magee. I have to go back all the time just to reread it. It
has everything in there.
Andrew:

Yeah. And final question, what is the best way for listeners to get in touch with you?

Ralph:

Probably sending an email through the LSPindexes.com site. There is a contact page on there and if you
went to www.LSPindexes.com you can reach me that way. That will get forwarded on to me wherever I
am at that time.

Andrew:

That's great. So thanks so much for your time today Ralph. Is there anything else you'd like to mention
before we wrap up today?

Ralph:

No, thank you very much Andrew and hopefully I will get the chance to meet you in person in the next
year or so.

Andrew:

Yeah, come on out to Australia.

Ralph:

Will do.

Andrew:

Just don't come in the wintertime. [Laughs]

Ralph:

[laughs] okay!

Andrew:

So thank you again for your time today, I know you are busy man. I highly recommend everyone to get
Ralph's books. I will have links to those in the show notes for this episode so yeah, thanks again today
Ralph I appreciate it.

Ralph:

You too Andrew I enjoyed it.

Andrew:

Cheers!

Ralph:

Bye-bye.
***End***

Ralph Vince Interview www.bettersystemtrader.com

Page 18 of 18