Abacus Fund Management LLC Presentation

1
THIS COMPOSITE PERFORMANCE RECORD IS HYPOTHETICAL
AND THESE TRADING ADVISORS HAVE NOT TRADED
TOGETHER IN THE MANNER SHOWN IN THE COMPOSITE.
HYPOTHETICAL PERFORMANCE RESULTS HAVE MANY
INHERENT LIMITATIONS, SOME OF WHICH ARE DESCRIBED
BELOW. NO REPRESENTATION IS BEING MADE THAT ANY
MULTI-ADVISOR MANAGED ACCOUNT OR POOL WILL OR IS
LIKELY TO ACHIEVE A COMPOSITE PERFORMANCE RECORD
SIMILAR TO THAT SHOWN. IN FACT, THERE ARE
FREQUENTLY SHARP DIFFERENCES BETWEEN A
HYPOTHETICAL COMPOSITE PERFORMANCE RECORD AND THE
ACTUAL RECORD SUBSEQUENTLY ACHIEVED.
2
ONE OF THE LIMITATIONS OF A HYPOTHETICAL
COMPOSITE PERFORMANCE RECORD IS THAT DECISIONS
RELATING TO THE SELECTION OF TRADING ADVISORS AND
THE ALLOCATION OF ASSETS AMONG THOSE TRADING
ADVISORS WERE MADE WITH THE BENEFIT OF HINDSIGHT
BASED UPON THE HISTORICAL RATES OF RETURN OF THE
SELECTED TRADING ADVISORS. THEREFORE, COMPOSITE
PERFORMANCE RECORDS INVARIABLY SHOW POSITIVE
RATES OF RETURN. ANOTHER INHERENT LIMITATION ON
THESE RESULTS IS THAT THE ALLOCATION DECISIONS
REFLECTED IN THE PERFORMANCE RECORD WERE NOT MADE
UNDER ACTUAL MARKET CONDITIONS AND, THEREFORE,
CANNOT COMPLETELY ACCOUNT FOR THE IMPACT OF
FINANCIAL RISK IN ACTUAL TRADING. FURTHERMORE,
THE COMPOSITE PERFORMANCE RECORD MAY BE DISTORTED
BECAUSE THE ALLOCATION OF ASSETS CHANGES FROM
TIME TO TIME AND THESE ADJUSTMENTS ARE NOT
REFLECTED IN THE COMPOSITE.
3
VK-Alpha CTA Selection Model
4
VK-Alpha Introduction

• More than 10 years ago, while working as Director
  of Managed Futures for a major international
  banking organization, Emil van Essen, who is now
  the President of VanKar Trading, observed that
  many of the investment funds that had allocated
  capital to a portfolio of the top performing
  CTAs of the day were paradoxically
  generating sub par performance. 
• After much testing, he discovered that most such
  CTA portfolios were constructed with heavy
  reliance on managers with high Sharpe ratios, low
  correlation of returns and lengthy track
  records.  One thing, however, seemed clear the
  then-current methodology did not work. 
  Regardless of how successful the CTAs had been
  to date, how much due diligence was performed and
  how thorough the analysis of each CTAs results,
  the performance of the CTA portfolio was, at
  best, unremarkable.  And yet even today, many
  years later, most funds are still constructed in
  exactly the same way.

5
VK-Alpha Introduction

• If traditional measures are poor predictors of
  future CTA performance, then how can an investor
  identify who next years star performers will
  be?  Perhaps more importantly, how can investors
  avoid those CTAs who will collapse and disappear
  in a sea of red ink? 
• After many years of work at solving this complex
  puzzle, VanKar and its team of professional
  software developers, engineers and mathematicians
  have made a significant breakthrough which allows
  us to combine CTA-specific performance data with
  an analysis of general market conditions in a
  model that predicts who the best and worst
  performing CTAs will be for the next 12
  months.   In test after test, the model has
  demonstrated an uncanny tendency to consistently
  outperform both our competitors and the major
  market benchmarks.  In fact, no other known model
  even approaches the consistent performance of the
  VK-Alpha CTA Selection model.

6
Background

• According to the National Futures Association,
  there are currently over 1,700 registered CTAs
  (including those who are dual registered as
  Futures Commission Merchants, Introducing Brokers
  and/or Commodity Pool Operators).  Thoroughly
  reviewing the strategy profiles and performance
  returns of each of these manager candidates could
  take weeks or even months.  An investor would
  most likely find that each prospective manager
  offers its own unique trading methodology that
  promises extraordinary results but does not
  provide an accurate basis for assessment versus
  its peers.  So how do investors decide which
  CTAs will help them most effectively meet their
  investment goals?

7
Background

• In response to this dilemma, the VK-Alpha CTA
  Selection technology, a sophisticated analytical
  model that carefully collects and reviews
  performance data for over 400 CTAs with a wide
  variety of trading models and strategies, and
  then selects a portfolio of those CTAs
  appropriate for the strategic investment
  guidelines of the investor or fund.   The
  VK-Alpha technology focuses on CTAs with proven
  track records, strong risk management controls
  and low correlation to each other.  Our tests
  have shown that an investor who uses our VK-Alpha
  technology can reasonably expect a significant
  long term advantage over other alternative
  investment portfolios.
• At VanKar, we set out to build a
  quantitatively-driven model that provides
  investors with the technology necessary to
  achieve their portfolio goals.

8
VK-Alpha CTA Selection Technology

• The VK-Alpha CTA Selection technology was
  developed over a period of ten years using
  meticulous quantitative analysis of the
  performance data from over 400 CTAs.  VanKar has
  identified a number of CTA performance
  characteristics that have proven to be excellent
  predictors of future CTA success.  Under rigorous
  test conditions, a portfolio selected using the
  VK-Alpha CTA Selection technology has
  historically shown significantly higher risk
  adjusted returns and low correlation to familiar
  market benchmarks.
• Many investors select their CTAs using methods
  that are both outdated and proven to be
  unreliable predictors of investment success,
  e.g., high risk adjusted returns, high Sharpe
  ratio, recent performance strength and
  non-correlation to the other investments in their
  portfolio.  Our groundbreaking research and
  state-of-the-art analytics have resulted in the
  VK-Alpha CTA Selection technology, which allows
  us to predict and select the CTAs that we feel
  have the greatest prospects for success.  We are
  confident that our methods, the product of years
  of intensive research and development, will help
  Abacus outperform the universe of Managed Futures
  programs and other alternative investment
  vehicles. 

9
The Database

• Most CTAs report their trading results on a
  monthly basis.  We have built an enormous
  database that includes over 20 years of such
  results from most active CTAs in the Managed
  Futures industry.  As we collect the data, it is
  first filtered to eliminate those CTAs with
  insufficient track records and/or negative
  lifetime returns.  We then normalize the data so
  that the returns may be viewed in risk-equivalent
  terms. 
• These adjusted results are then entered into the
  VK-Alpha CTA Selection model where a number of
  additional calculations and tests are
  administered.  Ultimately, the model produces for
  each CTA a fingerprint that summarizes that
  CTAs key characteristics.  These characteristics
  are both manager-specific (performance, risk,
  markets traded, etc.) and general economic
  (interest rate levels, overall market volatility
  by markets, etc.). 

10
The CTA Rankings

• The VK-Alpha CTA Selection model compares the
  fingerprint of each CTA to those of all the other
  CTAs at key points in history.  Once we have
  reviewed these historical matches, we can
  determine which CTAs are most likely to show
  strong performance over the next 12 months.  We
  are then able to rank them from highest to lowest
  based on expected risk-adjusted returns.   These
  quantitatively-based rankings act as the
  foundation of our ability to create portfolios of
  CTAs that match the requirements of our clients.

11
VK-Alpha CTA Selection Model
Formatting Data Standard performance database
is categorized and pre-formatted for use with our
software.
12
CTA Selection Overview
Portfolio Analysis Our software contains the
model. Using historical data, we can simulate
the growth of 1000 if a portfolio was
constructed and managed using the model.
13
CTA Selection Overview
Portfolio Analysis The software also tells us
current rankings of all investments so that we
may prepare for future allocations.
14
CTA Selection Overview

• Here we can see under "Money Allocation Methods"
  that we are allocating equal amounts of equity to
  10 different investments to the highest ranked
  CTAs.
• In addition to the model, we also show the
  average monthly percent return for the "market"
  which is the average of all qualified
  investments.
• The market return shows us the result if we had
  invested in all available investments.

15
CTA Selection Overview

• Our model portfolio is constructed using the top
  "x" ranked investments in each re-allocation
  period.
• Instead of picking the top rank, what if we
  picked the bottom rank?
• The return labeled "0_10" shows us the portfolio
  result if we had picked the investments that
  ranked in the bottom 10 as ranked by our model.

16
CTA Selection Overview

• Does the model rankings have any predictive
  power? If it does, a portfolio constructed of
  low ranking investment at the time should show
  below average future results.
• Portfolio of high ranking investments should
  predict above average future results and mid
  ranking investments should show average future
  returns
• Continued on next slide

17
CTA Selection Overview

• That is exactly what happens. Below are
  portfolio results based on each decile of
  predicted model rank and their corresponding
  portfolio future returns.
• Obviously, the prediction is not exact but the
  model rank is directly related to the future
  portfolio return.

18
CTA Selection Overview
Model Reporting Once a simulation is run, a
report can be generated to determine the makeup
of the portfolio at each re-allocation period..
19
CTA Selection Overview
Data Export The report may be exported to
excel for further analysis.
20
The Technology at Work

• Our work in selecting an appropriate portfolio of
  CTAs is the product of in-depth discussions with
  our clients in order to determine their
  requirements with respect to performance and
  risk.  Therefore, there is no standard VanKar
  portfolio all portfolios are customized and
  unique to the client.   However, for purposes of
  our demonstration here, let us assume that we are
  working with two different clients and that each
  of these clients has different investment needs. 
  
• Client A desires exposure to Managed Futures, but
  does not want to actively switch from one CTA to
  another in order to maximize returns.  Client A
  desires a more passive management style and we
  will, therefore, recommend and structure a more
  conservative approach.  Suppose further that our
  Client B wants to be more active in switching
  CTAs in order to achieve the highest possible
  returns while still holding the CTAs for a
  reasonable period of time.  We have applied the
  VK-Alpha CTA Selection technology independently
  to the needs of the two clients and developed a
  different hypothetical approach for both of
  them.  The two hypothetical approaches are
  identified below as (A) Standard and (B)
  Aggressive. 

21

The Technology at Work

• Each of the two approaches shown below is based
  on a hypothetical portfolio of ten actual CTAs,
  although in practice the number of CTAs selected
  will, of course, vary depending on the needs of
  the client.  Both approaches employ the same
  power to predict, but vary primarily in how often
  CTAs are switched in the portfolio.  Generally,
  the more aggressive the switching, the better the
  long term results. The approach selected will
  also depend, of course, on the investment
  parameters of the investor.  The VK-Alpha CTA
  Selection model strives to match an investors
  goals with a selection approach that has the
  highest probability of meeting their risk and
  return needs

22
CTA Selection Overview

• Model A and Model B are two models that have
  been created by VK-Alpha.
• Model A is more conservative with a long holding
  period while Model B is more aggressive with a
  shorter holding period.
• Figure 1

23
CTA Selection Overview

• Figure 1 demonstrates the normalized growth of an
  initial theoretical 1,000 investment from
  January 1995, until today, a period of over 10
  years, using each of the two approaches outlined
  above.  Perhaps the first thing that Figure 1
  highlights is that the returns of both
  approaches greatly exceed that of the Stark 300
  Index, a performance benchmark based on the
  returns of over 300 managers from the CTA
  community.  Looking closer, Figure 1 also shows
  that the returns for the aggressive approach
  outstrip those of the standard approach.
• The aggressive approach allows the client to
  remain regularly invested with the CTAs ranked
  in only the top percentiles.  The average holding
  time for each CTA in the portfolio using the
  aggressive approach is, on average, two and a
  half years.  The average holding time for each
  CTA in the portfolio under the standard approach
  is, on average, approximately four years. 

24
Figure 2 demonstrates the return of both our
hypothetical approaches over a 10 year period.
Again, that the results of the Abacus CTA
Selection model outperform the Stark 300 Index
benchmark over the same period is clearly
apparent.
25
CTA Selection Overview

• It is important to keep in mind that the VK-Alpha
  CTA Selection technology is constantly ingesting
  new data and performing iterative calculations
  for each of over 400 CTAs in order to determine
  which are most likely to outperform the others. 
  Because the technology is dynamic and the CTAs
  rankings change monthly, the aggressive approach
  allows the portfolio to regularily switch out of
  lower ranked CTAs and into the highest ranked
  CTAs.  This aggressive approach helps maintain a
  portfolio of very high ranked CTAs and therefore
  is likely to produce the highest possible future
  returns. The standard approach is still very
  successful, but sacrifices some return in order
  to maintain a longer hold period for the average
  CTA.  
• What is important to note in Figures 1 and 2
  above is the relationship between the two basic
  approaches.  The aggressive approach yields the
  higher returns over time but requires more
  frequent switches and, therefore, the most
  account maintenance.  The standard approach posts
  lower returns but requires less switching between
  CTAs.  Figures 1 and 2 also highlight the fact
  that the returns for both models greatly exceed
  the returns for more conventional measures such
  as the Stark 300 Index.

26
CTA Selection Overview

• Having examined the relationship between the
  various approaches elicited by the VL-Alpha CTA
  Selection technology, we now examine a current
  (December, 2005) snapshot of the model at work. 
  Figure 3, on the next page, demonstrates the
  year-to-date performance of those CTAs that the
  model has selected as the ten most likely to earn
  the highest returns.  For purposes of comparison,
  Figure 3 also includes the return over the same
  period for the Stark 300 Index and the mean
  average of the ten CTAs featured.  The two bars
  to the far right side of the Figure show that the
  mean return for the period is a positive 5.17 ,
  while the 300 CTAs featured in the Stark 300
  Index returned a positive 3.36 over the same
  period.  The remaining ten bars feature the
  individual performance history of those 10 CTAs
  that ranked highest in the monthly VL-Alpha CTA
  Selection rankings.

27
Figure 3
CTA Selection Overview
28
CTA Selection Overview

• On the other hand, Figure 4 on the next page
  demonstrates the year-to-date performance of
  those CTAs that the model has selected as the
  ten most likely to earn the lowest returns over
  the next year.  Again, as with Figure 3 above,
  Figure 4, next page, also shows the return over
  the same period for the Stark 300 Index and the
  mean average of the ten CTAs featured for
  purposes of comparison.

29
Figure 4
CTA Selection Overview
30
CTA Selection Overview

• Not all successful CTAs perform well all the
  time.  In both Figures 3 and 4,on previous pages,
  the CTAs selected are identified by name. 
  Figure 4 shows that there are some highly
  reputable managers with substantial assets under
  management who, for whatever reason, had poor
  performance during the testing period.  These are
  CTAs that, without the predictive power of the
  VK-Alpha CTA Selection technology, investors
  might have selected themselves and could easily
  have in their own portfolios today.  
• The VK-Alpha CTA Selection model forecasted
  that these CTAs would not perform well over the
  period and Figure 4 demonstrates that this
  prediction has indeed ultimately become a
  reality. In other words, the VK-Alpha CTA
  Selection technology would have helped investors
  avoid downturns in their own portfolio returns. 

31
Figure 5
CTA Selection Overview
32
CTA Selection Overview

• Figure 5 compares, from several
  perspectives, the year-to-date performance of the
  top-ranked CTAs with those of the bottom-ranked
  CTAs.  In each case and regardless of the point
  of comparison, the gap between the top-ranked
  CTAs and the bottom ranked CTAs is
  significant.  These tests confirm that the
  VK-Alpha CTA Selection technology can accurately
  predict which CTAs will outperform their peers. 
  In other words, our model does what it is
  supposed to do.
• As one might expect, the performance of
  the top picks is in all cases positive and the
  performance of those in the bottom picks is in
  all cases negative. Moreover, the magnitude of
  the gains of the CTAs in the top 20th percentile
  is greater than the magnitude of the gains of the
  CTAs in the top 50th percentile.  And, not
  surprisingly, the magnitude of their losses is
  smaller.

33
Summary of Results

• As shown in each Figure above, our test
  results demonstrate that the VK-Alpha CTA
  Selection technology successfully predicts CTA
  performance.  This allows investors to act with
  confidence when selecting CTAs and allocating
  assets.  Similarly, investors can avoid those
  CTAs with the lowest expectation of success. 
  The Figures above clearly demonstrate that the
  CTAs selected by the VK-Alpha CTA Selection
  technology have performed promisingly when
  compared both to their peers and to other market
  benchmarks.

34
Conclusion

• VanKar Trading offers a dynamic new approach
  to alternative investments.  Our VK-Alpha CTA
  Selection technology collects and reviews
  performance data for a broad universe of
  Commodity Trading Advisors, and structures
  portfolios of these managers that can provide
  investors with a significant long term advantage
  over traditional investment portfolios. 
• We believe that our technology provides the
  best risk/ reward potential available in
  alternative investments today.  For further
  information, please contact Emil van Essen at
  [phone]

35
What makes VanKar Different?

• Experience
• and
• VK-ALPHA

36

• Emil van Essen
• Over 17 Years of Financial Industry Experience.
• Original Director of Managed Futures with the
  BANK OF MONTREAL.
• Principal and President of the Asset Management
  Division of Vankar Trading Corp.
• A long standing registered Commodity Trading
  Advisor (CTA) who is also a component of the
  Stark 300 CTA Index.
• Prominent global keynote speaker on CTA Selection
  methods.
• Consults worldwide on the construction of
  multi-manager Alternative Investment Portfolios.
• Former Director of Quantitative Futures Analysis
  with Refco LLC.

37

• John Karvelas
• Over 15 Years of Financial Industry Experience.
• Principal President of the Brokerage and Trading
  Operations of VanKar Trading Corp.