This project is no longer maintained...Please
visit
two more simulated
Trading-by-Rotation webpages: Predictions
and TbR for Oil and the Euro and TbR with S&P500 and VIX
Hosted by the Department of
Economics, University of Peloponnese
Please read the Disclaimer
EurOil Index, Trading-by-Rotation
and TbR logos © 2007/2008 Dimitrios D. Thomakos
An exercise in Trading by Rotation |
|
Created and maintained by Dimitrios D. Thomakos, © 2008 |
|
Last updated: 03/06/2008 每 Weeks
running: 9 Trading updates available after Wed each week |
Figure on right
updated on: 03/06/2008 每 cumulative value of 1 Euro over past 325 weeks,
rolling window, rotation strategy
Visit two more simulated
Trading-by-Rotation webpages: Predictions
and TbR for Oil and the Euro and TbR with S&P500 and VIX
Click to link on this
webpage:
Summary of Trading
Recommendations,
Introduction, Data, Methodology,
Simulated
Trading and Performance
Latest news
03/06/2008 每 The Greek market has fallen dramatically last
week, over -7% for the FTSE-20 that the three stocks of the strategy belong to.
All three stocks had negative returns in excess of -5% and the cumulative
return of the strategy is at -25% - still, the historical performance of the
strategy is consistently better than either any of the three stocks or the
FTSE-20 index.
02/28/2008 每 Recovery!
The last recommendation for ETE was not optimal but both ETE and EFG went
into positive territory. The cumulative value of 1 Euro over the past 8 weeks is
now 0.8290, still outperforming the FTSE-20 index.
02/20/2008 每 Oops!!! A
bummer! The last recommendation was not good, at all...The recommendation
was for ETE that actually dropped over -6% compared to past week. This was a
hard blow, especianlly since the Greek market has been falling since the
beginning of the project. Nevertheless, we continue. The current recommendation
is still for ETE. Lets see what happens next week...
02/13/2008 每 A good
rotation! The latest recommendation for
01/30/2008 每 All three stocks bounced back to positive territory compared to past
week! Our directional forecast was for ETE, realized return 8,66%, but
01/17/2008 每 A sharp
drop! The sharp decline in European markets has affected Greek stocks and,
of course, the strategy. On 01/15 and 01/16 the ASE has dropped cumultatively
-6.9% and the FTSE-20 has dropped cumulatively -7.23%. This had strong effects
on the three banking stocks we are examining. As the strategy does not involve
cash the effect was quite pronounced, with the two-week cumulative value of 1
Euro falling to 0.8096. This sharp drop is evident at the end of the series in
the figure above.
Summary of Trading Recommendations
Trading recommendations are for the week of 03/12/2008 and
are based on computations and data of the previous week.
Evaluation Period in
weeks |
25 |
50 |
75 |
100 |
125 |
150 |
200 |
300 |
325 |
Recommended Asset |
ETE |
ETE |
ETE |
ETE |
ETE |
ETE |
ETE |
ETE |
ETE |
[1.] See the data section for more on the asset nomenclature. The
abbreviations correspond to three major Greek banks, EFG Eurobank ERGASIAS S.A. (EFG), PIRAEUS
Bank S.A. (PEI) and
National Bank
of Greece S.A. (ETE).
[2.] The asset in bold corresponds to the
evaluation period that has yielded the highest cumulative return in the
historical evaluation.
Welcome to the TbR webpage! Here you can find
a real-time application of a trading method based on asset rotation. This is a
follow-up extension to the EurOil Index project that I am also
maintaining, where I present results from rotation-based trading for Oil and
the Euro. Please visit the EurOil Index for additional information. In
this project I expand the rotation strategy to three, rather than two, assets
and I am using a different asset class and market. Specifically, I am using
stock returns for three major Greek commercial banks that are being publicly
traded in the Athens Stock Exchange. The choice of these three banks is based
on their market capitalization as they
are high-capitalization companies and are considered industry leaders in
the Greek banking system.
I am using data from publicly available, internet-based
sources. The electronic editions of many Greek newspapers give free online
access to historical data on companies that are listed in ASE. I collected the
data from the website of Eleftherotypia newspaper because it
provides the longest historical series. The three banks that I consider are: EFG Eurobank ERGASIAS S.A. (denoted EFG in the
figures), PIRAEUS
Bank S.A. (denoted
The trading
methodology I am using here is based on some of my past research. I use a
modification of the methods presented in the paper ※Market Timing and Cap Rotation§, by D. Thomakos, T. Wang and
J. Wu, Mathematical and Computer Modeling
(2007), vol. 46, pp. 278-291 (available on http://www.sciencedirect.com),
to construct a trading strategy that rotates between the three assets based on
their predicted signs. The methodology can be seen as a generalized &sign
forecasting* approach although it is probably more informative than simple
pairwise sign forecasting. The strategy gives long-only signals and does not
take into account transactions costs so the results presented should be
interpreted accordingly.
A
rotation-based strategy will heavily depend on the choice of assets used in the
rotation. For example, the weekly drawdown will always equal the minimum return
among the three assets, if that particular asset is (wrongly) selected.
Therefore the strategy*s success depends on whether it can correctly predict
the relative sign of the assets most of the time and note that we do not need
to be completely accurate all the time.
This is illustrated in the historical evaluation that I present in the next
section.
Simulated Trading and Performance (using data up to 03/05/2008)
Cumulative returns (based on longest
evaluation period of 325 weeks)
Trading recommendation for Wednesday (based on data a week
before) |
Asset |
Realized Return % EFG |
Realized Return % PEI |
Realized Return % ETE |
Success in Direction |
Cumulative Value of 1 Euro |
01/09/2008 |
ETE |
-7.08 |
-6.06 |
-0.21 |
+ |
0.9978 |
01/16/2008 |
PEI |
-13.19 |
-18.82 |
-11.68 |
- |
0.8096 |
01/23/2008 |
PEI |
-8.41 |
-3.79 |
-9.33 |
+ |
0.7717 |
01/30/2008 |
ETE |
6.66 |
9.81 |
8.66 |
- |
0.8583 |
|
EFG |
-1.78 |
-3.66 |
-1.22 |
- |
0.8405 |
02/13/2008 |
PEI |
-0.53 |
0.46 |
-4.06 |
+ |
0.8451 |
|
ETE |
-1.93 |
-3.59 |
-6.33 |
- |
0.7817 |
02/27/2008 |
ETE |
5.07 |
-0.09 |
4.73 |
- |
0.8290 |
03/05/2008 |
PEI |
-5.40 |
-7.90 |
-7.50 |
- |
0.7500 |
03/12/2008 |
ETE |
|
|
|
|
|
[1.] The realized return for each asset is the weekly return
(from Wednesday to Wednesday) from the week before to the week of the trade 每
return measured as difference in log prices.
Discussion (based latest results)
The
figures below summarize the trading performance of the proposed rotation-based
strategy vs. the performance of each of the three stocks and the FTSE-20 index,
in which the three banks participate. I present the accumulated value of 1 Euro
invested at the beginning of different evaluation periods (Figure 1) and the
corresponding information ratios (Figure 2). The strategy clearly outperforms all
three individual stocks for almost all evaluation horizons, trailing closely
the performance of PEI. The strategy also outperforms the FTSE-20 index for all
evaluation horizons as well. For an evaluation period of the last 100 weeks the
strategy yields 1.20 Euro vs. 0.87 Euro of ETE and 0.91 Euro for the FTSE-20.
For an evaluation period of 325 weeks the strategy yields 2.57 Euro vs. 1.67
Euro of PEI and 1.39 Euro for the FTSE-20.
In addition to its superior performance in terms of accumulated wealth,
the rotation strategy has the highest information ratio across all evaluation
periods.
To
gauge the frequency of correct (and incorrect) decision-making of the strategy
I calculated the average number of times that the strategy picked the asset
with the highest (lowest) return. Surprisingly, these averages did not change
drastically with the length of the evaluation period, something that can be
considered a sort of &robustness* property of the strategy. The average number
of correct decisions was about 40% (ranging from 35% to 44%) while the average
number of incorrect decisions was about 25% (ranging from 16% to 29%). Note
that this means that about 75% of the time the strategy did not pick the worst
(lowest return) asset.
Figure 1 (updated 03/06/2008)
Figure 2 (updated 03/06/2008)
Disclaimer: The contents of this webpage are provided for information purposes only. Prices shown in this webpage are indicative and the author is not offering to buy or sell or soliciting offers to buy or sell any financial instrument. The views in this publication are those of the author alone and are subject to change at any time, and he has no obligation to update the opinions or the information in this webpage. The author of this webpage does not accept any liability whatsoever for any direct or consequential loss arising from any use of the information provided. The information in this webpage is not intended to predict actual results, which may differ substantially from those presented. The author is not affiliated with any of the companies used in the analysis that is presented in this webpage.