ADVANCE SYSTEMS SUMMARY
My name is Ed Ayoub. I am a physicist working in the field of computer science. I have designed a comprehensive commodities trading system based on mathematical analysis that I have developed to improve fund management performance. I am seeking partners who are looking for a solid, mathematically-based system by which to hedge trades. Here is an overview on how the product functions.
OBJECTIVE:
The key objective of this venture is to utilize our proprietary, numerically-based system for the prediction of price movement. We seek partners with expertise in funding, fund management and investment, as we do not endeavor to manage funds ourselves. Our role in this venture is to operate the program that supplies the daily predictions. The role of trading funds will be assumed by our investors.
Our biggest strategic advantage is the uniqueness of the system, as it is improbable for other investors to have derived a system like ours. Our system should appeal to investors who want to diversify to hedge against risk, as it is based on solid data from the past performances of trades.
HISTORY:
I started writing the program when I was seeking patterns in the futures markets back in the 1980's. I tried different ideas on paper, with pencil & calculator. I started out by programming the calculator, and taking that output and comparing, recursively. I began to program the computer to take over this process, incrementally increasing its ability to generate its own formulas based on the new input of data. I did this for years. Eventually, I ended up with an artificial intelligence program specifically designed to locate patterns in the futures market. (Futures were chosen due to the large amount of data available on their daily activity).
HOW IT WORKS:
Our system uses the daily high, low and closing price of every future contract of all the commodities, indices, currencies and interest rate instruments. It outputs concrete buy, sell and liquidate advice of the most likely profitable trades. Not all trades advised will be profitable but the aggregate of all the advice generated yields an average of 40% profit annually, with a standard deviation of 20%, based on 20 years of back testing.
We are currently working on new hedging algorithm strategies. Currently, the only method of risk management is the diversity inherent in the system. We do not advise use of options or derivative instruments. Our system is designed to enter and exit all positions at market on close and does not utilize stop loss orders. It was designed as such so that execution of the trades at the desired price is nearly assured.
One thing I have learned is how easy it is to be fooled by statistics. With a small pool of data it is easy to be deceived in thinking that there are correlations when none exist. I have learned a number of ways to avoid this. The best is running the system again on never before used data. This is the main reason for deciding to use the same system for all items traded as futures: commodities, indexes, bonds or currencies. To derive the best system it is ideal to keep the amount of variables down and keep the pool of data as large and accessible as possible.
INVESTORS:
As mentioned earlier, we seek partners with expertise in funding, fund management and investment. We will depend, in part, on the recommendations of these partners for risk management, tax efficiency and hedging strategies. Our partners could best answer your questions on prior experiences, roles and responsibilities at previous positions, tax efficiency, and business risk. They would be best to designate who will monitor risk and what risks to mitigate.
BUSINESS STRUCTURE:
I am the sole author of the system. My brother serves as advisor. My intention is to continually develop the system, support it and market its output. Our growth potential depends on who we take on as partners. As previously mentioned, we will operate the system to supply the daily predictions and continue to maintain and develop the system. Other necessary business roles will be assumed by our partners.
We do not yet have a CFO, marketer, operations head, or analyst. Nor have we retained the services of lawyers, accountants, or prime brokers as of yet. It is currently undecided how staff members will be compensated but they would most likely be paid by commission and/or by salary and commission.
START UP COSTS:
I would expect to launch the fund with a trial of $100,000, and a range of 1-10 million dollars to start. I will supply quarterly reports and be prepared to answer all investor questions to the best of my ability. I will happily travel to set up satellite test systems for potential investors if need be, but I do not intend or wish to travel for marketing purposes. I do not have a lot of personal money to invest, however, it is my life’s work and its success is crucial to me. I am willing to do whatever it takes to get this system to function according to the goals presented herein.
I expect our new partners to determine if we will have any strategic investors, how we plan to resolve conflicts with other managed funds, as well as questions about fund structure. My guess is that we will target 15-30 investors at start up. We would be open to negotiation on the issue of side letters and leave this decision up to our initial investors. All transactions will be transparent. Lock-up of the fund should best be 18 months, with special benefit for longer lock-up.
FUND DETAILS
To ensure everything is working smoothly to the benefit of investors, we suggest a hurdle of 25% gain a year, a minimum individual investment size of $25,000, and management fees of 1% a year - plus 20% of profit from inception. We further suggest exposure and size limits to be 5 million per investor in the short term, and in the long term to be less than 10% of the total fund value. The diversity of the fund will vary with system predictions and averages as shown below. In particular, we consider 8 currencies, 6 indices, 3 bonds, and in the physical commodities, 12 agricultural or livestock products, 5 metals and 2 crude oils and derivatives.
At this time the only metrics to manage portfolio risk and types of hedging strategies is the number of these commodities. (We are working on a long term solution that involves a complex algorithm.) We will use the leverage of normally margined futures but in relation to the value of the total fund the leverage should never exceed much more than 100%, with possible exceptions in the case of interest rate instruments.
EXCEPTIONS:
Being that this is a new concept and that there may be capacity issues with some of the commodities we spread into, we would initially close the fund to new investors after the first 20 million dollars. However, because our system works well with currencies, we could, with some slight modifications, potentially accommodate investments of any size. We estimate that the limit may a percent of the total currency market. Within these restrictions we will gladly will work with all investors (including seed capital investors) agreeable to a minimum lock-up period of 18 months.
MARKET STRATEGY
As called for by the system we will initiate or liquidate positions market on close. This is our investment process. It handles both the long and short side. This system is how we generate investment ideas, it is our competitive advantage and how we will differentiate ourselves from others in this field. Though they are unlikely to have a similar product, we consider our competitors to be other quantitative investors. We do not rely on street research. The market data analyzed by our system is our only research tool. At this point, our hedging strategy is the diversity inherent in the system. We have drawn no correlations between our performance and market environment but the best period was in 2002 and the worst draw down has been in the period of July 1998 to October 1998.
TESTING
Our system is a predictive system. It is designed to produce a profitable outcome for any market traded in futures. We have run it on 50 of the most common (American) markets on which we possess data. If you have historical data for other items, we would be willing to set up a test station for you. We could set you up with a system, without seeing your data, and still remain confident that your results will be satisfactory. However, the system needs a minimum of 5 years back history. (Keep in mind the exclusions noted below - we need to discuss in advance whether your items would be good candidates for our system).
EXCLUSIONS
Our system is based solely on past data. It is important that this data truly represents the item traded. If an item has very high volume and open interest in an exchange, we can assume that the blueprint of the item is truly represented by the daily data. If not, we believe that the probability of noise is higher and the results are not predictable.
For these reasons we eliminated the following from our portfolios: Feeder Cattle (FC), Lean Hogs (LH), Orange Juice (OJ), Pork Bellies (PB) and Rough Rice (RR).
Another cause of distortions in the data is possible spread trades with other items of higher volume and open interest. For these reasons we excluded the Dow Jones Index (DJ), which correlates to the S&P 500; the One Month Libor (EM) and Federal Funds (FF), which correlate to the Eurodollar; Heating Oil (HO) and Propane (PN), which correlate to Crude Oil (CL). However, we have cautiously decided to keep the Nasdaq 100 (ND), the Midcap Dow 400 (MD) and the Russell Index (RL). We have also thrown out the Nikkei Dow 225 (NK) and the Euro Yen (EY), since we would expect the data from their home exchanges to be more relevant.
Lastly, the price of natural gas is about as volatile as the gas itself! It is easily the most volatile future traded, capable of doubling or halving its price in hours. We have always included natural gas’ historical data in the development of our system. However, we can not recommend currently trades in it due to these levels of volatility. For this reason we excluded Natural Gas (NG).
With these exclusions we show a net gain in 34 of the 38 traded futures, with a modest loss on Cocoa (CC) and Sugar#2 (SB). Since our system requires a minimum of 5 years historical data before it generates predictions, the Euro and Peso have not yet had a sufficient trial. To date, while the Euro has had a majority of positive trades with our system, it shows a slight loss. This should moderate in the near term, and we expect accurate predictions on both these currencies. This may not the case for Sugar or Cocoa. At this time we do not have an explanation for their poorer performance over our 20-year tracking period. Though we do not have an explanation, we have included them, (along with the Euro and the Peso), and they simply the statistical losers in our system.
As the system is designed to predict the movement of the cash price, the results are slightly better when trading the actual cash value of the instrument (when possible) as opposed to the futures. In either case, the annualized returns are 33% to 43%. (The standard deviations vary slightly, as do the corresponding Sharpe ratios). Our sample fund below will trade in the most liquid market of the particular item. In currencies, we will trade in the cash market. In most other items, we will trade in the futures market.
SAMPLE FUND PERFORMANCE
Statistics for the 20-year time period from 1/1/1987 to 1/1/2007 below are for a simulated cash fund. (Statistics for a simulated futures fund are in parentheses when available). The fund makes single futures contact trades according to the system’s prediction. There is no compounding that occurs in these figures. The reason for expanded growth is largely due to the opening of new markets in recent years.
DRAWDOWNS
Previous Peak
Margin
Drawdown
percent of last surge up
Time Period
$43,598.45
$19,729.89
$46,031.24
106%
from 2/21/1988 to 10/26/1990
$268,081.49
$24,591.88
$92,785.58
34%
from 7/19/1998 to 10/ 9/1998
$331,850.79
$20,894.38
$61,069.38
39%
from 1/23/2000 to 1/29/2000
$365,592.73
$22,293.82
$59,272.41
63%
from 3/26/2000 to 10/26/2000
$418,633.58
$19,428.64
$51,735.64
46%
from 8/19/2001 to 9/22/2001
$450,796.08
$21,264.36
$41,251.36
49%
from 10/28/2001 to 10/31/2001
$554,236.58
$11,730.04
$51,167.54
35%
from 8/25/2002 to 10/10/2002
$574,692.18
$9,754.39
$40,666.74
57%
from 8/31/2003 to 8/14/2004
*Final Value at 12-31-2006: $611,211.08 ( $631,061.08 without the $50 commission per trade)
STATISTICS
Sharpe Ratio
1.74 (1.65 )
without margin or commissions
1.86
Annualized Gain
$30,560.55 ($29,507.61)
Standard Deviation
$15,266.14 ($17,148.42 )
Annualized Gain
32.94%
Standard Deviation
16.45%
Hurdle
21.6%
Hurdle Rate
65.0%
Best Year - 2002:
116.9%
Worst Year - 1998:
-16.5%
CORRELATIONS & BETA
Index
Correlation
Beta
GSCI Commodity Index
-0.041
-0.014
CRB Commodity Index
0.076
0.201
S&P 500
0.216
0.183
NASDAQ 100
0.187
0.102
RISK MANAGEMENT
The primary focus of this program is the accuracy of predictions. Very little attention has been made as to hedging and risk minimization. There is, however, some hedging inherent in the program. See statistics below:
ALLOCATIONS
ITEMS
OTH
OIL
MET
$$$
IDX
INT
COM
C&$
C$I
ALL
M&$
percentage of time hedged with another in category
24%
0%
8%
6%
6%
2%
35%
41%
52%
56%
15%
percentage of positions in this category
20%
3%
10%
7%
55%
5%
33%
40%
95%
100%
17%
percentage of time this category is long
58%
53%
79%
72%
97%
65%
64%
66%
84%
83%
76%
LEGEND:
OIL= crude oil and derivatives MET = metals OTH = any other commodity $$$=currencies IDX= indices INT= interest rate instruments COM= all commodities C&$= all commodities and currencies C$I= everything except interest rate instruments ALL=all items M&$=metals and currencies
(Note that a position in the Dollar Index (DX) is considered long if it is short and short if it is long, for the sake of accurate comparison)
* Our fund will trade or invest in both the long and short side of commodities, stock indices, currencies and interest rate instruments. The gross and net exposure will vary but, on average, we are on the long side 83% of the time, with an average of 4 positions or a total of 0 to 9 at any given time.
For comparison’s sake only, the same statistics of the all 50 items would be:
DRAWDOWNS
Previous Peak
Margin
Drawdown
percent of last surge up
Time Period
$42893.45
$22129.25
$ 70851.60
165%
from 2/21/1988 to 10/26/1990
$250084.19
$ 21392.67
$ 96715.77
35%
from 7/19/1998 to 9/ 1/1998
$271921.19
$ 37280.51
$ 85571.81
72%
from 8/ 8/1999 to 10/21/1999
$399252.33
$ 66146.54
$ 109125.83
51%
from 3/26/2000 to 2/ 6/2001
$560920.08
$ 42655.95
$ 102488.15
38%
from 8/31/2003 to 8/ 3/2005
$568429.58
$ 56127.49
$ 70978.49
49%
from 7/ 5/2006 to 9/ 1/2006
*Final Value at 12-31-2006: $595,900.88 ( $621,450.88 without the $50 commission per trade)
STATISTICS
Sharpe Ratio
1.36 (1.13 )
without margin or commisions
1.50
Annualized Gain
$30,347.24 ($21916.77)
Standard Deviation
$18,848.51 ($18,756.29)
Annualized Gain
27.81%
Standard Deviation
17.27%
Hurdle
18.3%
Hurdle Rate
55.0%
Best Year - 2002:
99.2%
Worst Year - 2004:
-20.0%
CORRELATIONS & BETA
Index
Correlation
Beta
GSCI Commodity Index
-0.033
-0.083
CRB Commodity Index
0.066
0.213
S&P 500
0.309
0.322
NASDAQ 100
0.425
0.284
RISK MANAGEMENT
The primary focus of this program is the accuracy of predictions. Very little attention has been made as to hedging and risk minimization. There is, however, some hedging inherent in the program. See statistics below:
ALLOCATIONS
ITEMS
OTH
OIL
MET
$$$
IDX
INT
COM
C&$
C$I
ALL
M&$
% of time hedged w/ another in category
29%
3%
8%
5%
11%
13%
41%
48%
62%
68%
15%
% of positions in this category
23%
5%
9%
7%
50%
6%
38%
44%
94%
100%
15%
% of time this category is long
61%
51%
79%
72%
95%
55%
64%
65%
81%
80%
76%
LEGEND:
OIL= crude oil and derivatives MET = metals OTH = any other commodity $$$=currencies IDX= indices INT= interest rate instruments COM= all commodities C&$= all commodities and currencies C$I= everything except interest rate instruments ALL=all items M&$=metals and currencies
(Note that a position in the Dollar Index (DX) is considered long if it is short and short if it is long, for the sake of accurate comparison)
* As above, our fund will trade or invest in both the long and short side with varying exposure. In this example we have an average of 5 positions or a total from 0 to 11 at any given time.
CONCLUSION
We are continually developing our system. These statistics are for our most recent version of Advance Commodity Trading System.
I hope this outline answers all or most of your questions regarding our system. Please let me know if I can expand on any topic for your further review.
Thank you for your time.
Sincerely,
Edward G. Ayoub
19 Hanson Street
Somerville, MA 02143
(617) 625 2823
BACKGROUND:
Over 25 years of experience in Software Engineering and Systems Development
Areas of expertise include: Digital Image Processing, Machine Vision, Automated Systems, Instrumentation, Device Drivers, Graphical User Interface, Mathematical Modeling / Algorithm
I have been working on this project extensively for the last 25 years
EDUCATION:
M.S. Physics, 1988, University of Massachusetts, Amherst, MA
B.S. Physics, 1979, Worcester Polytechnic Institute, Worcester, MA