Money Management 7

back to

Money Management 6

Money Management:

The best way to test this (an interesting experiment for your thesis) is to
try a positive expectation game. Have 10 marbles 6 black (winners) 4 white
(losers); randomly select and return the marbles 100 times. Have room full
of people trying different money management systems starting with a
hypothetical $100,000; winning or loosing what they risk for each "trade".
(Or record the sequence and do it yourself). The results will range from 0
to Millions. Try a 70% game. The results will also range from 0 to millions.
This is why the worlds best and largest traders have very simple systems but
sophisticated MM rules.

The best definition I've come across for MM is "how many". Your system
should tell you when and where to get out, only after these decisions have
been made should you then use your MM system to decide HOW MANY, but the two
decisions should not over-lap.

You are right to try antimartingale systems, increase with wins, decrease
wth losses. The exact nature of the system determines the best formula. With
wide stops like an n-day breakout, something like 2% is a good starting point, however
this would be suicidal with close stops. A volatility based formula would
stop be better to stop you getting too carried away.

> I also tried to increase/decrease position size according to the
>dependency among the last x trading days of a trade. When there is a
>Win-Lose-Win-Lose pattern I increase-decrease-increase-decrease, when
>there is a Win-Win-Win-Win pattern I increase continuously (e.g. every 5th
>day). Also not very satisfactory.
>Am I on the right track but on the wrong branch?

Close.
Example 1: If your MM rules are to risk 2% of a $100,000 account ($2,000)
Your system stop is $1000 away so you take 2 positions. The market moves up
and you have a $4,000 profit but your system exit has not moved. You are
now risking giving back 4% ($4,000) If you want to keep a constant 2% risk
you may consider reducing risk back to 2% and just selling one at the
market.

Example 2: The same as above except your system has a reason (a technical
one, not a MM one), to move up the exit stop. The new stop is risking only
1% now ($1,000) so you could now double your position size a buy two more at
the market. With this MM system you could increase dramatically your
position size while never risking more than the original 2% to the market.

Example 2 is the basis of the very successful turtle trading system along
with a very simple n-day channel breakout entry and exit system. There are
many variations on this i.e. you could risk say 1% for initial positions and
4% marked to market or MARKET HEAT as it's sometimes called.

The idea is to limit risk and maximise profits by letting your tested
technical system give you a reason to enter and exit the market, and MM to
say "HOW MANY" but the two are quite separate decisions.

Andrew Dykes at "Kerry O'Connor" koad@ihug.co.nz


Money Management/Expectancy/Expected Value:

>1) I tried a very close system stop, e.g. a $500 stop while Volatility (in
>terms of a 15 day Average True Range) was higher than $500 at least in 2/5
>of the cases. - I was surprised as the results were not too bad, as I could
>"pack" more Nr. of Contracts with a close stops which obviously paid for
>the higher number of stopped out positions.
>Do you consider such close stops as hazardious?

As long as your realise that this $500 stop is not a money management rule
but a trading rule. If you wish to include it in your system then by all
means test it and see what it does to your expectancy (!) ***. A MM rule does not
alter your expectancy therefore is not at trading rule.
>
>2) What I was thinking about is to have a recursive money management system.
>That means I just look at what results a certain system produces, e.g. an
>average loss of $1000, an an average profit of $3000 and has tendency to win
>big after a loosing streak and vice versa (lose bigger after a winning streak).

I've not found this to be true to any meaningful extent. I would caution against trying use so called "streaks" to determine outcome. This is the very psychological bias that allows casinos to prosper. If you know the expectancy (!) *** of your system then every trade has the same expectancy going in, the actual result of any one trade is close to random. Also I've found the only way to have the "real world" trading to match up to your hypothetical expectancy is to trade multiple markets and use many years of data in your testing.

>Any ideas how to build the average loss/average profit into a money
>management plan? - Unfortunately I never played Black Jack, otherwise I
>would maybe have an idea.
>To know to suffer a loss in 60% of the time, and to have a profit in
>40% of the time is advantage (Expected Value = 60% x -$1000 + 40% x $3000 =
>$600), but I am not quite sure how to implement it into a valid method.

To see how good a system is I like to then divide the expected value $600 by
the average loss ie $600/$600 = 1 Which means your expectancy is $1.00 per
dollar risked. This is a great way to compare different systems.

If the system has been tested properly with adequate slippage an expectancy
of .5 is tradeable and .7 is a good one. An expectancy of 1.0 will make you
rich!

Here is one plan, however the unfortunate truth is only with a large account
can one truly use it or any effective MM plan,

Firstly I use my gambling formulas to determine maximum bet size for
greatest return (unfortunately this will result in large drawdowns and you
have to be psychologically prepared to trade through these.)

I use the Kelly Criterion. Kelly % = A-[(1-A)/B]
{A is the % of winning trades in decimal format (reliability of system) and
B is the average profitable trade in $ divided by the average losing trade in $

e.g. a coin flip game: the reliability of the system is 0.5. In this game you
make twice what you risk when you win. Thus B = 2. Therefore the amount of
remaining equity you should risk to produce the maximum rate of return is

Kelly % = 0.5 - [1 - 0.5)/2] = 0.5 [0.5 / 2] = 0.5 - 0.25 = 0.25

The maximum bet then is 25%.

Now with this information I would take 80% of Kelly to be on the
conservative side (20%) I then work out how many different markets I will be
trading i.e. 10. So I would allocate 2% of my trading equity per position.
If in this example my account is $100,000. For each new position I would
subtract my initial entry price from my exit price then determine "How
many". I might be able to afford 4 corn (if the amount risked was $500 or
less) only 1 Swiss Franc and maybe have to forgo an S&P if the initial stop
was further away than $2000.

If I have on 10 positions and all of them go against me on one day I'm
still within the Kelly Criterion. But I know the "how many" has not changed
my expectancy (!) *** in any way.

>Is this what I call a recursive money management system? You decide AFTER a
>trade what (rather 'how many') to risk next? This obviously requires
>dependency between the trades. But I also read very often one should see
>each trade independently, what a contradiction.

No, I have never seen any evidence of dependency between trades. A well
thought out system should have the entry and exit point fixed before the
trade is entered. The difference is your risk. The money management
discussion "how many" is then based on your total account size.

>The problem I had (to transfer the experiences from the game to the futures
>markets) was a) that I did not know if I should consider daily changes (daily
settlement) as a streak or the last x closed out trades (which could be weeks away).

There are 3 basic approaches:
Core Equity Method: to use only the value of your account from closed out
trades to determine positions size, less the amount risked on each open position.

Total Equity Method: the cash in your account as above, plus the value of
any open positions.

Reduced Total Equity Method: A combination of the two. It's like the core equity method but you add back open profits if the stop has been moved into profit.

>b) that I often increased risk when the market was already lost steam, so
>big drawdowns were the result.

The only way to avoid this is to trade many un-correlated markets at once.
The selection of your portfolio is crucial. (And not an easy task by any means).

>What do you mean by 2%? - Initial risk?

Yes 2-3% of your account size for the initial risk and maybe double that for market heat (Marked to market on the close). However this market heat should not mean moving your stops (this would change the expectancy (!) *** and would mean you are trading a different system) but by varying position size. This is of course the luxury of a large account and is not usually possible for the individual trader staring out. In fact it doesn't necessarily increase overall profitability but should produce a smoother equity curve, and cause less psychological problems.

I can't recommend highly enough the "Special Report on Money Management" by Van Tharp (I mentioned it in my last e-mail) which covers all of the above in great detail (different position sizing methods, different models of money management).

>> For each new position I would subtract my initial entry price from my
>> exitprice then determine "How many".
>You are probably talking about the stop loss (exit) point here or do you use target exits
>(at a profit)?

Yes stop loss exits. I never use profit targets. I always follow the basic cut losses short, let profits run. The reason this works is in the nature of distribution of the trades. There is not the classic bell shaped curve but one with "fat tails" i.e. there are more large moves than would be expected if they were random. By having an initial stop to prevent the unusually large adverse moves, and a trailing stop to attempt to capture as much of the extraordinary large moves, it is possible to profit in the long run.
As a result all profitable systems I've seen are trend following and have less than 50% win to loss ratios.

>Do you try to be in different markets all the time (i) or is it possible
>that you trade only one at the time (ii)?

When I say I'm in the market this includes buy stop orders above the market or sell order below the market. I then "trade" about 35 market world-wide and would have open positions in about 60% of them at any one time.

It's all about having a positive expectancy then applying it as much as possible. As a blackjack player I used a system which had the odds of one player coming out ahead over a weekend to 2 out of 3. Two players together could, by combining banks effectively double the number of days and the probability of winning would be 3 out of 4. By the time we had a dozen players we won every day!
Trading is the same, the more markets the better. I would even add a market
if I knew it was less profitable than the others (of corse this is impossible to know in advance) because the smoother my equity curve the higher my bet size can be.

>If (i), you must have a very 'easy-going' Entry Signal, which gets you in
>very often, otherwise you couldn't be in, let's say, 5 markets simultaneously.

As I said previously having an order in the market is the same to me as being in it. But don't be too worried about entries, it's a common beginners fallacy. It's the same as those "Lotto" type lotteries where you get to pick your own numbers (birthdays, lucky numbers etc..) Obviously as an educated Man you know this makes no difference to the outcome.
Entries are the same. If you don't believe me try using all the well known
entry techniques, Channel breakout, moving average cross-overs, volatility entry etc and if they are profitable after 5,10,20 days few are even as good as random entry.
They seem important because the decision when to get in is the last time a trader is in control. Exits are at the whim of the market. Only money management makes any real difference on your profitability.

>>Is this what I call a recursive money management system? You decide AFTER a
>>trade what (rather 'how many') to risk next? This obviously requires
>>dependency between the trades. But I also read very often one should see
>>each trade independently, what a contradiction.

I know of no dependency between my trades. "How many" is a function of
account size and the percentage you are willing to risk.

>Do you measure correlation among markets with the LN (Natural Log) of price
>changes? What I even find difficult here is how do the programming efficiently (I do
>all the calculations with MS Excel): I must have all the data (of all commodities involved)
>open to calculate a correlation matrix in a certain time window. This eats up all my
>memory... Maybe I should calculate the correlations beforehand and write the
>results into a separate file where it is read from during simulation? Danger is to 'lose >contact to reality' here. Which program do you use to run your calculations?

None of the above. I don't consider correlation among markets to be a mathematically problem, rather a common sense one. For example I'm currently short Swiss Franc Futures so I don't have a position in D-Marks. Just from looking at the charts of these markets they obviously move in a similar fashion. From experience I know to sell the weaker contract and buy the stronger. Grains and precious metals obviously have some correlation as do
bonds and Euro's. S&P's have a negative correlation to bonds but not always. It is just a matter of paring down position size when you feel you may be over exposed to a particular sector. I know of no formular which improves on experience in this matter.
I know this is not too helpful to an academic study of trading however the nature of market a forever changing. Oil and Gold moved together in 70's but by the 80's they were quite different.

The point to remember when testing a trading system, is that testing is not about accuracy or exactness. We only have last years data to test on when we really need next years! It would be similar to trying to find out who would win the next Austrian Elections by polling the French, i.e the wrong data pool.

To quote William Eckhart "The delicate tests that statisticians use to squeeze significance out of marginal data have no place in trading. We need blunt statistical instruments, robust techniques. I.e a robust statistical estimator is one that is not perturbed by mistaken assumptions about the nature of the distribution."

>I have heard about the fact that moving stops would change the expectancy and varying
>position size will not. But what is the (mathematical) explanation of this?

>I do not really understand why a money management rule does not alter my
>expectancy.

Your expectancy is per dollar risked. A 2-1 expectancy game does not change
if the bet size changes.

>Is it correct that any additional rule which alters my expectancy (up or
>down) is dangerous and should be avoided?

It is important not to have too many degrees of freedom. Do not use rules which would only affect a very rare occurrence. If I had to choose between two otherwise equal systems I would go for the one with the lessor number of imputs or degrees of freedom. This would be the one that would be less "curve fitted" and more likely to work in the future.

>I know the Kelly formula for quite a while but never tested it.

The kelly formula should be use only for total portfolio risk. Assuming the worse case scenario and all positions act as one and move against you at the same time. It is much too large for an individual position.

>In practical terms, that means you decrease position size during the life of a trade instead
>of moving your stops closer, right?

Yes, although this depends on the system. The only way to know is to test all possible
variations. Remember trading is a much psychological as mathematical. Over all profitability is irrelevant if the trader cannot stomach the drawdowns.
So changing position size during the life of a trade is more to do with keeping a constant risk but doesn't necessarily increase profitability. If your exits are particularly clever it may be better to keep the bet size constant.


Andrew Dykes at "Kerry O'Connor" koad@ihug.co.nz

Fixed Ratio money management technique (Ryan Jones' company):

As you stated, Optimal F is the quickest way mathematically to increase your
account balance. As you also noted, it comes at the expense of large drawdowns.
Additionally, better systems will typically have larger Opt. F drawdowns because
better systems yield a higher optimal fraction, such that when a loss does occur,
you have a large amount of your account at risk.

Ryan's "Fixed Ratio" money management formula looks to maintain the drawdown of an
account at a constant percentage level. In other words you increase or decrease
the number of contracts traded in order to keep your drawdown at roughly the same
percentage level. This theoretically makes the account easier to trade because you
don't have the wild equity swings associated with optimal f. This money management
approach boosts account returns well above single contract trading, but still
falls short of optimal f levels. It's just at matter of whether or not you could
withstand (both psychologically and in regards to account capital) a 60% or 90%
drawdown that is quite typical of optimal f.

I've purchased and worked with Performance I (Ryan Jones' software package for
fixed ratio). At $1200, the software leaves something to be desired (pretty much a
glorified spreadsheet) although a new version was just released (free to current
owners). Basically what you're paying for is the concept, which is fully
disclosed. Bottom line: Fixed Ratio is unique, but not stellar. I believe if you
study Vince's material (opt. f) and Kelly's techniques and fixed fractional
trading and get an understanding of how money management works, you could probably
come up with an approach that works just as well as Fixed Ratio. If you're looking
for someplace to spend money, however, this is as good a place as any!

Neil Peplinski qa3135@email.mot.com

Never risk more than one days range. In general, it is preferable to exit in the direction that the market is going I believe. I am also interested in this type of discussion, because I have a great entry technique and am more interested now in a great exit technique. Most trading systems expound on great entries but ignore how to have great exits. Many Gurus recommend scaling out of trades in order to cut down risk. Finally, when the risk is greatly reduced and some profits are protected, a final contract is left to go for the big long home run. The first few contracts are taken off at targets, and the final one is let run with a loosely trailing stop.
For small accounts, a profit target is probably preferable for short term systems and a trailing stop is preferable for long term systems. The system that I am trading allows your profits to run for one day, and then you nail them the next day. If the market keeps going in your
favor, you reenter. The system skips a lot of profit between where you took profits and reenter. I am thinking about how to fill the gap.

Money Management-Van Tharp:


Long trades have a high positive expectancy but make you money only 25% of the time. When the market starts up trending (winning streak on the long side), the probability of it continuing is about 70%. A winning trade could last 15 days or more, and (if you risk the maximum allowed) you could double your equity each day. Thus, a winning trade that was allowed to double each day for 15 days could pay more than 16,000 to 1. However, if it didn't go 16 days, you would lose all of your profits. For more information on expectancy or probabilities, see the questions and answers below and Dr. Tharp's comments on money management.

Short trades have a negative expectancy but they make money for you 75% of the time. However, you could loose up to 20 times the amount you risked. Thus, any time you risk more than 5% of your equity on the short side you risk bankruptcy. A losing trade only
lasts one day, but a winning trade can go as long as 15 days and as short as two days. The art form to winning this game is to be able to capture as much profit as possible from a winning trade while, as the same time, not letting the profit get away. For example, if a winning trade went 15 days, you could make as much as 16,000 times your initial risk if you continually risked the maximum. But, if you continually risk as much as possible on a
winning trade, you will eventually lose all of your profits.
All trades are not necessarily 1:1. The market can move in your favor or against you at 1:1, 2:1, 5:1, and even as much as 20:1 on the short side. There are occasional 2:1 losses on the long side.

During the course of play, each player has a listing of the last 10 trades they placed. Here is how to interpret some of the columns:

Risk - the amount risked during the trade
L/S - indicates whether the player traded on the Long side or the Short side
Mkt - indicates the market movement:
Up = a win on the long side or a loss on the short side
Down = a loss on the long side or a win on the short side
Chg - indicates the amount the market moved
Close - your closing equity (Equity + Amount)


YOUR LAST 10 TRADES

Date Equity Item Traded Risk L/S Mkt Chg Amount Close

06/12/1996 50000 Ethiopean Rice 00500 L DN 01 -00500 49500
06/25/1996 49500 ABC Shoes 00880 L DN 01 -00880 48620
07/03/1996 48620 Monitors Inc 00200 L UP 02 +00400 49020
08/28/1996 49020 Monitors Inc 00200 S UP 01 -00200 48820
09/01/1996 48820 Manual Trans 00200 S DN 01 +00200 49020
09/04/1996 49020 Cups Mugs Inc 01000 L DN 02 -02000 47020

Thus, if you risked $1000 on the long side, as it did with "Cups Mugs Inc." and the market moved down by 2 times, then you would end up losing $2000.

If a player elects to "go short" and the market enters a long winning trade, you won't be permitted to participate (except by going short) until a loss occurs and a new item is presented. The reason for this is that we are simulating the whole issue of allowing profits to run, but not letting them get away from you. By going short, you have elected not to participate in that particular trade. Remember that a winning long trade can last many
days. Thus, when you go short, you may be locked out of the game (except to go short again) for a number of days.

Only one stock or commodity is traded per day. During a winning streak, the stock or commodity being traded remains the same from one day to the next. Once a loss occurs, a new stock or commodity is offered. During the winning streak, you can risk a maximum of your initial trade on that item plus any profits you have realized since the item was offered. If you risk $0 on the long side, you will not be able to participate in a winning trade because the amount you can risk is limited to $0 throughout the trade.

Players that are unable to return to the Virtual Trading Exchange daily are letting their profits ride (once they have some) just like in the real markets. However, we do have an error in the
logic of the game in that program only lets your profits ride during the second day of your participation in a winning streak, but not after that. In fact, if you let your profits ride, by not
participating, you risk getting the results of the trade on the day you return (instead of the next day). We will not fix errors of this type, so you are on your own if you try it. You need to
take this into consideration when considering your money management prior to placing your last trade, and then place the trade accordingly. You can also decide to place a "stop" in the
market by entering a "0" as your last trade, effectively preserving your capital.

There may be circumstances where you will notice two trades listed on a single day. This tells you that the exchange automatically executed your trade at the end of a winning streak (which would have occurred during one of the days which you were not participating but had a trade on) before switching to a new item. By not participating, you were letting your profits ride. This is the only circumstance which will cause two trades to be listed on the same day.

When a winning trade starts, it may last as long as a week or two. It is considered to be just one trade. If you elected to take it on the short side or risk $0 on the long side, you have
essentially decided not to participate in that trade on the long side. However, you can make more short entries on that trade - should you decide that the winning streak should be over. Your risk on the short side is only 1:1, but remember that a winning trade can last a long time.

There is a lot of uncertainty in the markets, just like there is a lot of uncertainty in the game. In other instructions, we say that certain things will probably happen. However, nothing is
certain. And if the unexpected happens, it is your problem. We will not resolve such problems in your favor. It's one of the hazards of trading.

How to Play the Simulation:

The best way to describe how to play the simulation is to take you through a series of trades to see what will happen. The names of the stocks or commodities used in this example are purely fictitious - any resemblance to actual companies is purely coincidental.

Let's say that the first trade is MicroLoft, Inc. On day 1 you place a trade for $1,000 on the long side. The next day, you look to see what happened in the market and find that the
market moved against you at 1:1 and you've lost the $1,000 (if you had gone on the short side, you would have won).

TIP: Each day, the market switches instruments unless there is a profit on the long side.

Now, on day 2 everyone is trading on Gold. You place a trade for $1,000 on the long side. The next day, you look to see what happened in the market and find that the market moved against you again, this time at 2:1 and you've lost the $2,000 (if you had gone on the short side, you would have won again).

On day 3, everyone is trading Silver. Again, you place a trade for $1,000 on the long side. The next day, you look to see what happened in the market and find that the market moved against you again, at 1:1 and you've lost another $1,000.

Your new equity is now $46,000. If you had gone on the short side for all three trades, your new equity would have been $54,000.

On day 4, everyone is trading WXYZ Corp. Again, you place a trade for $1,000 on the long side. The next day, you look to see what happened in the market and find that the market moved in your favor at 1:1, and you've now won $1,000. However, if you had decided to go short on day 4, you would have lost at 5:1, or $5,000, and would not be permitted to trade on the long side until a new instrument is offered.

___________________________________


TIP: When there is a win on the long side, the short side can lose anything from 1:1 to 20:1. Once the market enters an up trend, it enters a new mode of play. WXYZ Corp. is now trending up. It will remain the instrument offered as long as it remains in the up trend
(i.e.: until the first loss). Subsequent, days are just a test for you (those of you who've made money on the long side) to be able to let you profits run, but also lock in some of the profits by raising your stop (i.e., not risking the maximum).

On day 5, WXYZ Corp is still being traded. Your decision now is whether to risk all your winnings or to raise your stops. You decide to trade $2,000 on the long side. The next day you look and find the market is still up trending and you have won at 1:1, or $2,000. You have now made $3,000 on this trade!

On day 6, WXYZ Corp is still being traded. You decide to transfer $500 into your equity by only trading $2,500 today. The next day, you look to see what happened in the market and find that it moved in your favor, and you won $2,500.

Your new equity is now $51,500 with $5,000 to risk.

On day 7, WXYZ Corp is still being traded. You decide to move an additional $1000 into your equity account by only trading $4,000 today. The next day, you look to see what happened in the market and find that it moved in your favor, and you won $4,000.

Your new equity is now $55,500 with $8,000 to risk.

On day 8, WXYZ Corp is still being traded. You decide to only risk $2,000 on the trade. The next day, you look to see what happened in the market and find that it moved against you,
at 1:1. This trade is now concluded and you have $53,500 in equity. You've made a total of $7,500 on WXYZ.

On day 9, Arizona Oceanfronts is being traded and the entire process begins again.

It pays to cut your losses short and to let your profits run. You do this by implementing good money management techniques. Our simulation has been designed to provide you interactive feedback during the course of play. The simulator has been programmed with a set of rules. If it senses that you are trading in a highly risky manner, you will receive
an instructional email message to help you out.
___________________________________

Questions & Answers:

I don't have access to the internet over the weekends. Why does the simulation run continuously?

The simulation runs continuously by design. During the design phase, we considered the problem that some people that may only have internet access at their office during "regular business hours". However, our simulation is being played by people all over the world in various time zones, thus it was impossible for us to define what "regular business hours" would be for each player.

Furthermore, the goal of the Virtual Trading Exchange is to teach people money management skills. People need to learn to consciously consider their money management strategies each time they place a trade. Thus, if a member is unable to participate over a weekend, holiday, business trip or any other time, then they need to consider this when deciding how much to risk on the last day that they are able to participate. Their money management should tell them to either reduce the amount they are willing to risk on the trade or to place their trade on hold by placing a stop in the market (entering a "0") as their last trade.

This game seems like gambling, not trading, to me. You'll even lose on long trades most of the time. What am I suppose to learn?

It's not gambling because, unlike the casinos, the game has been programmed to give you a positive expectancy if you go long on every trade. It is actually a good simulation of long-term position trading. The losses only last a day and you've got to let your profits run. You also have to learn to make sure that your profits don't get away from you. Thus, if you play the game well, you'll learn the lessons necessary to be a good long-term trader/investor.

I don't have a lot of information about any particular stock or commodity that you give. How am I suppose to make a decision?

Actually, taking every trade in this simulation is a little like following a system. You don't know which trades will make money, but you know that you will make money if you take them all and manage your money well. You also know that you have a positive expectancy by trading on the long side.

The key to making money in the markets is money management. That is what we are emphasizing in this game. Most people think that the key to success is analyzing the market and finding the next winner. That's why most people have a lot of trouble making money. We're trying to teach you how to make money in the markets through money management. Hopefully, you'll be open to learning about it.

How can I make a decision on this trade (it could be any one) when I can't see a chart of it?

The answer to this question is again the same as the last one. Follow the expectancy and go long. The key to this game is money management (how much to risk) and how much to continue to ride on a trade during a winning streak. Going long or short shouldn't be a part of your decision making. However, we allow you that option because most people like to make stupid mistakes.

Why not give us a portfolio to trade?

A portfolio would be useful, but our goal with this simulation is to teach you money management. That's easier to do when you only have one item to concentrate on. We want to make the lessons simple and easy for everyone to understand.

What's a good strategy to use in this game?

We don't want to give one strategy and have everyone play that. However, we will give you some general hints:

Trade a percentage of your equity on each trade --- enough to do well, but not so much that you'll get too far "in the hole". You might consider increasing the percentage when you are ahead and decreasing the percentage of equity when you are behind.

When the market is up trending (a winning streak on the long side), conserve some of your profits. On method is by protecting a fixed percentage of your profits (which is like using a trailing stop in the market). When the market starts up trending, you may only protect a small percentage at first, but as the trend continues, protect more of your profits. One way you might trail stops (and there are many) would be to trade 80% of the maximum on day 2; 60% of the maximum on day 3; 40% on day 4; and 20% on day 5 onward. However, you can do any number of possibilities in terms of protecting your stops. For example, if you risked 100% of what's possible on days 2 and 3 of a winning streak, you'd risk losing it all, but you'd have an 8:1 (2^3) = (((1/0,25)^3)^0,5) = (((1/0,25)^(3*0,5)) winner after three straight winning days. If you risked the maximum in a trade that went ten days, you would have a 512 (2^9) (((1/0,25)^9)^0,5) to one winner. However, in order to get that profit, you would have to risk all of your profits on days 2 through 10 - which is probably not a wise decision.

Incidentally, you can determine how much you are allowed to bet during a winning streak by risking your total equity. The computer will reject the trade and telling you how much you can
risk. You can then elect to trade a percentage of that figure.

I am having trouble reconciling the following statements: Long trades have a high positive expectancy but make you money only 25% of the time. Short trades have a negative expectancy but they make money for you 75% of the time. How can you have positive expectancy with only a 25% success rate, but negative expectancy with a 75% success rate?

___________________________________

Expectancy is a function of both the probability of winning and the size of the average gain to the average loss. For example, imagine having a bag of marbles in which there are 99 black and one white. The 99 black ones will cause you to lose whatever you risk. The one white one will pay off 1,000 times what you risk. Each marble is replaced in the bag after it is drawn out. Thus, on any given trial, you only have a 1% chance of winning, but a great positive expectancy. I'd play that game all day, but you must be in the game to capture the 1000 to one marble. [1000 * 1% - 1 * 99% = 10 - 0,99 = 9,01]

Most people don't understand expectancy, so that's why we do it that way. To understand expectancy better I'd suggest that you either subscribe to our Market Mastery newsletter or purchase our Money Management Report. Also see the expectancy section of this web site.

When you say that their are only 25% winning trades, please look at the following scenario and tell me if there are 25% winning trades.

Gold Loss
ABC Stores Loss
Golf Shoes Win
Golf Shoes Win
Golf Shoes Loss
Copper Loss

Would you agree that there are only four trades there three losses and one winner that eventually loses. Thus, the winning percentage is 25%.

Yes, that is correct. There are three losing trades -- gold, ABC Stores, and Copper -- and one winning trade -- Golf Shoes. Thus, there are 25% winning trades.

Be looked at the trials in this game, and there is no positive expectancy. Your engineers need to re-program this thing.

I agree that we have some bugs in the program. The biggest bug is that it doesn't force you to let your profits run beyond the second trial of a winning streak. However, you have to take
advantage of those streaks (by letting your profits run) and you can't keep betting it all (which means you have to raise your stop). When you do that there are some big winners. For example, there was a potential 96:1 winner on the long side in one of the last two games (see below). Not one person took advantage of it. In addition, the random generation has not had any really big winning streaks in the past, but there is a good chance of one in the future.

There is only a positive expectancy on the long side when you allow your profits to run. For example, a good strategy after a win might be to risk 80% of the maximum on the second bet of a win streak; 70% of the maximum on the third trial; 50% of the maximum on the fourth trial,; 30 % of the maximum on the fifth trial and 20% thereafter. If the streak, only starts out with a one to one winner, might not want to risk more than 50% on the next trial.

In the last game, there was one trial that started out as a five to one winner. If you had risked a thousand on the first trial and let the whole winnings go for the five winning trials, you would have made $96,000. But let's say you did something like the following:
Trial Risk Amount Won New max. risk
1 $1,000 $5,000 $6,000
2 $5,000 $5,000 $11,000
3 $7,000 $7,000 $18,000
4 $9,000 $9,000 $27,000
5 $8,000 $8,000 $35,000
6 $6,000 ($6,000) $29,000 (Total Gain)

Look what happens, you make $29,000 out of that streak on an initial risk of $1,000. That's almost a 30 to 1 gain. Even if you left your risk at the initial win of $5,000, you would still have made 20:1. However, no one was willing to do that. And, of course, no game has a positive expectancy if you don't play the big marbles. Remember that the core risk on this trade - the risk to your core equity - was never more than $1,000.
During the rest of it, you were just risking open profits.

Some winning trades pay 5:1, others pay 2:1, and others pay 1:1.
What is the distribution of those trades? Also some long trades lose at 5:1, 2:1, and 1:1. What is the distribution of those?

If we've programmed things correctly for the current game (beginning September 21st 1997), the long trades can only win or lose at 1 to 1, with an occasional 2:1 loser on the first trade
only. (There has already been one exception to this on November 7th). The way you'll make a lot of money is to be able to as much as double your capital everyday by riding a winning streak 10 to 15 days. But of course, if you continually double your risk, you'll eventually lose it all. You have to lock in some of your profits, just like you would with a stop order in the
market.

I have made some simulated runs assuming a 25% chance of a winner being long and a 70% chance of a winner continuing the next day. I also assumed that a long wins and loses by 1:1 each day. I'm betting a constant size on each day, and stay with the initial bet during the
whole long trade if it trends. The bet size is $500. I recorded 100 runs and found out on the average one breaks even. To conclude, your rules don't give a positive expectancy on the long side.

The long trade is one trade. Your initial risk might be $500, but you can risk $1,000 on day two, $2,000 on day three, $4,000 on day 3, $8,000 on day four etc. By day 14, you could risk
as much as $446,000. Your core equity risk would still only be $500 since it is the same trade. Please understand that -- if ABC Stores wins for 8 straight days it is still only one trade. You can let profits ride in a trade. That's what is wrong with your calculations. Money management determines how much you risk initially. But money management also controls your position size during the trade. The golden rule of trading is let your profits run.

You need to lock in some of your profits. Otherwise, you'll eventually lose your core equity risk and all your accumulated profits in that trade), but you could theoretically make 16,384
to 1 on the initial risk of a trade that goes 15 days. A trade that goes 15 days is not that probably, but some big winners are.

What if I haven't played for a few days, how do I know if we're in a winning trade or not? And if we are in a winning trade, can I jump into the middle?

There is no way that you can know if we are in a winning trade or not if you've been away for a few days. However, the odds are much greater than the 25% that we will be in one if you've
been away for a while. Yes, you can jump right in. Just realize that if it is already going, it might not last that much longer.

I've been following the winners and number 26 appeared to go short and then go back in on the long side. The rules say that cannot be done and I haven't been able to do it. What gives?

That's a great question and it illustrates how one of the querks in the game, previously given in the instructions, works. When you leave the game for a period of time, the computer does
not process your trade until you return. If the trade is no longer in a winning streak when you return, it will be counted as a loser (even those it would have been a winner had you played the next day. Here's what happened with number 26.

11/06/1997, 0190815,Viper Technologies , 0001120, L, W, 01,
0001120, 0191935,
11/06/1997, 0191935,Viper Technologies , 0002240, L, L, 01,
-002240, 0189695,
11/09/1997, 0189695,Basket Ball Inc , 0001000, L, W, 01, 000
1000, 0190695,

Viper Technologies did not actually have a losing day until 12/7. However, #25 made a trade on 11/6 and then did not return until 11/9. Since the 11/6 trade had not been resolved, it was declared a winner, but the next trade was carried over and lost at double the best size. As a result, it might have looked like #26 really went short on 11/8 in Basket Ball Inc.. But was really happened was that several trading days were skipped.
__________________________________

If you look at what people are doing with money management in this game, it is really atrocious. First, if you want to get into the top 10 in any of the first three games you could probably bet $50 long on every trade and make it. Sometimes there are only 10 profitable people in a game (aside from our staff members and people who are cheating and playing two accounts at one time), and you get into the top ten if you make $500. Most of you want to get into the top ten immediately. Resist that temptation and just follow your discipline. Most of the people in the top 10 end up doing something foolish and then falling behind. Be patient.
Most of you give into the temptation to bet with the probability and you do so big time. Several people risked everything on the first trial and went bankrupt. If you risk more than 5% on the short side on any trial, you risk bankruptcy.

Percentage risk and percentage return do not have a symmetric effect on the portfolio, because the required percentage return to recoup a given loss (in percent) increases geometrically with the size of the loss. Not adjusting the trading size to the portfolio value (by decreasing with greater risk) would lead to an increase of risk at an increasing rate (as expressed by the increasing leverage of a fixed-size trading position).

The process of translating portfolio risk into number of units traded, under given price risk estimates, is shown here using a T-Bond position as an example.

Example A: Proposed LONG position in US T-Bonds at 100
(Unit Face Value US$ 100,000)
Estimated (Accepted) Price Risk: 2%
Allocated Equity to this trade: US$ 1,000,000
Accepted Portfolio Risk on Position: 1.5%
Calculating Accepted Position Risk in absolute US$ = (1,000,000 * 1.5%) = 15,000
Calculating Accepted Per Unit Risk in absolute US$ = (100,000 * 2% ) = 2,000

Accepted size of the trading position = 15,000 / 2,000 = 7.5 units of T-Bonds at face value US$ 100,000 equalling a US$ 750,000 investment. If contract sizes do not permit the exact trade size as calculated, we mostly round down to the lower possible trading size (lower risk), here, a US$ 700,000 investment.

Example B: Actual Market Price is now at 101 (up from 100)
System calculates price risk to be 3% (up from 2%).
Allocated equity now at US$ 800,000 (due to losses in other markets within the portfolio)
Accepted Portfolio Risk unchanged at 1.5%

Calculating Accepted Position Risk in absolute US$ = (800,000 * 1.5%) = 12,000
Calculating Accepted Per Unit Risk in absolute US$ = (101,000 * 3% ) = 3,030

Calculated accepted position size = 12,000 / 3,030 = 3.96 units of T-Bonds (at face value US$ 100,000), rounded up to 4 units, equals accepted position size of US$ 400,000. The system will therefore issue a signal to sell US$ 300,000 worth of T-Bonds in order to adjust the position to changes in the price risk and in the portfolio composition, expressed in the lower portfolio allocation.

CAPITALIZATION

Capitalization (actually, the lack thereof) is perhaps the most overlooked, misunderstood, and MAIN reason why so many traders starting out fail to achieve their goals. They go into their new business without an adequate supply of a traders lifeblood: CAPITAL. Think of yourself as a general or field marshal. You are fighting a war. If the war goes badly at first and your opening battles all end in defeat, then you must have adequate reserves to carry on the struggle until you emerge victorious. You lose all your soldiers, guess what? YOU LOSE!!! A general can not hope to fight a successful campaign without enough soldiers in his army. A trader can not trade his account properly if financial armageddon rides upon every trade.

YOU MUST BE SUFFICIENTLY CAPITALIZED! IF YOU ARE NOT YOU ARE DOOMED TO FAILURE FROM THE BEGINNING!

THE RULE OF 4.5

What then is sufficient capital to begin trading the S&P/DJ? As a common sense guideline I have developed a simple and straightforward method to determine what your MINIMUM account balance should be to open your account and begin trading. I call this THE RULE OF 4.5. Simply put, you can NOT risk more than 4.5% of your total equity on any single trade. This gives you the advantage of being able to lose 20 times in a row without having your account wiped out. (If you do lose 20 times in a row I want you to do 2 things: 1) Realize that you were not meant to be a trader. 2) Send me your methodology so I can fade it).

All kidding aside, there is something known as LUCK and STREAKS. If you begin trading 2 S&P contracts with $5000 and run into a bad streak (say 4 out of 5 losers for a loss of $2500) at the very start, you are going to seriously think about quiting right away! Thats a 50% drawdown in equity in a very short period of time...maybe even ONE day! It will be very difficult from a psychological standpoint to trade properly after that or have confidence in your abilities. Yet if you were properly capitalized you would have only suffered a %12 drawdown on those S&P contracts, not nice, but you can dig in your heels and continue the fight!
Below is a simple table to determine what your MINIMUM opening account balance should be based upon the Rule of 4.5. The amount of the contracts varies based upon the account balance. Remember: you must always begin with TWO contracts at a minimum, or you do not stand a ghost of a chance to be profitable.

OPENING ACCOUNT EQUITY: 7000 10000 13000 16000
# Dow Jones (DJ) you can trade: 2 3 4 5

OPENING ACCOUNT EQUITY: 20000 30000 40000
# S&P 500 you can trade: 2 3 4


If you stick to the above table, and do not deviate from it (and for goodness sakes, if you do deviate, do so on the side of caution), then you have an excellent chance of making money. If you disregard this advice, you do so at your own peril! THIS MAY WELL BE THE SINGLE MOST IMPORTANT ADVICE I AM GIVING YOU. IF YOU LEARN NOTHING ELSE, THEN LEARN THIS.

go to

Money Management 8

^{back to top}