|
StormGuard-Armor is
a composite of three measures we
call Price Trend, Market Momentum, and Value Sentiment.
Its improvements
do not come from trying to improve predictive
timing, but rather from actual event detection. These
are "tells"
 indicating behavior changes of momentum and
value investors. The Market Momentum and Value Sentiment
signals are extracted from market volume and new highs/lows data sources. Twelve separate measures of the Price Trend, Market
Momentum, and Value Sentiment signals are combined
using the methods of "fuzzy
logic" to
produce the final StormGuard-Armor value. Even when applied to the S&P 500
(right), the improvement is clear. |
|
StormGuard-Armor
Incorporates Event Detection, not Simply Timing
Adjustments.
Most market sentiment
indicators incorporate only measures of price change
using one or more moving averages. Although there are
many kinds of moving averages and many time periods over
which they can be calculated, it is often the case that
adjusting them to improve performance over one period of
time results in reduced performance in others.
Reacting more quickly generally results in additional whipsaw
losses, while reacting more slowly can result in going further
over the cliff during a true market crash.
The daily price record is limited in the amount of
helpful information it can provide.
In order to improve
performance, additional information from other sources is
required. Fortunately there are numerous
measures of other market characteristics available, including daily
volume, new highs and lows, advancers versus decliners,
and volatility. The objective is to find additional predictive
information that can be used like "poker tells" to improve investment decisions.
 StormGuard-Armor
is the composite of three measures we call Price Trend,
Market Momentum, and Value Sentiment.
The Price Trend signal is
essentially
identical to the original StormGuard,
which was derived from the application of Matched Filter
Theory and performs measurably better than the classic
50/200
Death Cross.
The Market Momentum signal further
incorporates volume information, and may be thought of
as a tell, indicating that momentum traders are
changing their behavior. Note that a
Wall Street "momentum trader" defines momentum as daily
volume times the rate of price change in much the
same way as a physicist defines momentum as mass times
velocity. However, ordinary trend followers generally mix the terms trend and
momentum together by inherently
ignoring the consideration of volume.
The Value Sentiment
signal incorporates new highs and lows information, and
may be thought of as a tell, indicating that value
investors are changing their behavior.
It
is important to note that the new Market Momentum and
Value Sentiment
signals are not just timing adjustments
 made in
hindsight, but actually provide signals prior to
serious price movements.
The first chart illustrates that during the market's rough patch
in 2010,
it was only the momentum signal that indicated trouble
in the market. In 2011, during the run up to the August
1 selloff triggered by the S&P downgrade of U.S. debt, the
Market Momentum signal again indicated trouble ahead of the
event. However, in the
second chart it is the Value Sentiment indicator that
indicated trouble long before the August 2015
market drop triggered by trouble in the Chinese markets.
The improvements
provided by StormGuard-Armor do not come about from
trying to improve predictive timing, but rather from
actual event detection. These are tells indicating
behavior changes of momentum and value investors. The
Market Momentum and Value Sentiment signals are extracted from the
additional data sources using
PID (proportional, integral, and differential)
control algorithms to condition them appropriately. The
Price Trend, Market Momentum, and Value Sentiment signals are then
combined together using the principles of "fuzzy
logic" to produce the final
StormGuard-Armor result.
|
SG-Armor+
Upgrade (Dec.2018)
• Whipsaw Loss Reduction: Upside Breakout Pattern
• Drop-N-Pop False Trigger: Glitch Recovery Pattern
 Whipsaw Loss arises when a precipitous
drop occurs before StormGuard generates a sell signal
and then subsequently returns to higher levels before
a new buy signal is generated. There have been six prior
instances of whipsaw losses during market corrections
over our 30-year database. All six instances are now
properly detected. Whipsaw loss has been either
significantly reduced or eliminated for all six
instances without any collateral damage. Although
StormGuard-Armor is currently at risk of another such
whipsaw event, SG-Armor+ is now in place to deal with it
effectively.
Drop-N-Pop False Triggers occur when a
fearful news event causes a precipitous, but quickly
reversed, market drop that is sufficient to trigger
StormGuard-Armor even though the event has already
reversed itself.
When triggering after the short event has come and gone,
the strategy is forced to sit on the sidelines for a
month while a bull market goes on without it. There have
been 3 prior instances of Drop-N-Pop False Trigger
losses over our 30-year database.
|
|
The Animated
Chart (above right - hover over to enlarge)
examines the most recent five years to illustrate
exactly when and how improvements were made in this
interval by detecting and acting on these short-term
market patterns. The animated chart cycles through four
slides: (1) Using the well-known
Death Cross
indicator (50-day average crossing the 200-day average)
to exit the market to a money market fund ($CASH); (2)
Using StormGuard-Armor to exit the market to $CASH; (3)
Using StormGuard-Armor+ to exit the market to $CASH; and
(4) Using StormGuard-Armor+ to exit the market to
the integrated Bear
Market Strategy BMS-A. Correction of the Nov.
2014 and June 2016 Drop-N-Pop false trigger events by
StormGuard Armor+ is easily noted, as is the whipsaw
reduction provided in March of 2016 and January of 2019.
Other similar events in our 30-year database are
likewise improved.
That devil in the details really matters.
How They Work: Both
upgrades utilize short-term pattern recognition
algorithms that were able to handle all identifiable events in the
30-year database without generating false triggers or
creating other collateral damage. There are no special rules
involving specific events. The Drop-N-Pop algorithm
disables StormGuard-Armor from triggering at month-end
if in the past 6 weeks there has been a 5% or more drop
followed by at least a 75% recovery by month-end during
a bull market. Whipsaw loss reduction is provided by
promptly detecting a satisfactory market correction
rebound during a bull market that reliably indicates the
market will continue heading higher. A rebound of at
least 5% with no prior higher-high in 7 weeks (or
slightly shorter periods for stronger rebounds) is the
test. When this condition is met,
StormGuard-Armor will indicate a fixed value of 0.02%
for at least 4 weeks, after StormGuard may again be
triggered providing one of the following conditions has
first occurred; (1) the StormGuard-Armor indicator
formally moves to a value above 0.025, (2) the
StormGuard Std. value becomes negative and is descending, or (3)
after 8 weeks the StormGuard Std. value has been in
decline for the last month.
Enabling StormGuard-Armor+: SG-Armor+ will
automatically be applied to all Strategies employing
StormGuard-Armor from 12/25/2018 forward
without the need to do anything. Any Strategy
that is substantially edited (more than its name or
display options) and posts a Last Edit Date of
12/25/2018 or thereafter will additionally apply the SG-Armor+ model
across the entire data set and will indicate its use of
SG-Armor+ in the chart's subtitle. You may experimentally disable SG-Armor+ by adding the
character ''-'' to the end of your Strategy name. You
must also further substantially edit the Strategy to
cause it to recalculate the entire trading history from
scratch. Similarly, if you remove the ''-'' you must
also substantially edit the Strategy for the change to
take effect across the entire trading history. You
must also further substantially edit the Strategy to
cause it to recalculate the entire trading history from
scratch. Similarly, if you remove the ''-'' you must
also substantially edit the Strategy for the change to
take effect across the entire trading history.
Application Notes: When you apply SG-Armor+ there are
some odd things that can happen, including (1) If you
use a Bear Market Strategy that performs well during one
of the improvement periods, then SG-Armor+ may provide
little improvement for that period. (2) Similarly, if
the trend leader picked by your Strategy performs poorly
(relative to the Bear Market Strategy) during one of the
"improved periods", then implementing SG-Armor+ may
slightly reduce the Strategy's performance. A
high-performance Bear Market Strategy that rivals your
Strategy's bull market performance during this
transitional period is not necessarily a bad thing, and
may simply indicate that the responsible fund is a bit
on the volatile side.
|
Best Bear Market Performance? Here's Proof:
The charts below illustrate the improvement
provided by StormGuard-Armor over (1) owning
only the trend leader, (2) Betterment's 60/40
stocks/bonds portfolio, or (3) simply owning
the S&P 500 index. Click the charts for
details. To reasonably compare AlphaDroid's
performance to Betterment and the S&P 500, a
Strategy was constructed holding only six
ordinary largecap and midcap ETFs (SPY, SPYV,
SPYG, MDY, MDYV, MDYG) along with the aggregate
bond ETF AGG. The Strategy for owning only the
trend leader has StormGuard disabled and the
algorithm simply selects the best trending of
the seven ETFs at the start of each month. When
StormGuard-Armor is enabled, the Strategy is
restricted to owning AGG during bear markets.
While the 60/40 stocks/bonds split helps reduce
Betterment's Max Drawdown and marginally
improves its Sharpe Ratio (a measure of average
return/volatility), it does so at the cost of
lower returns.
|
|
Market
Sentiment/Direction Indicator Comparative Performance
In the table and charts
below, the performance of seven market direction
indicators is evaluated for each of three major
classifications of ETFs (broadly diversified, U.S. sectors,
and world regions). The market direction indicator
determines whether it is a bull market (risk-on) or a
bear market (risk-off). During a bull market the
Strategy selects the trend leader to own at the end of
each month from among the Strategy's candidate ETFs.
During a bear market, the Strategy will either move to
the safety of CASH or alternatively own a long-term
treasury ETF, depending on its configuration.
Performance for both of these Bear Market Strategy
configurations is detailed below. Please review the
Bear Market Strategies page for a
discussion on how hindsight selection bias can be
avoided during bear markets using a well-designed
Strategy that selects from a diverse set of asset
classes that often (but not always) do well during a
market crash. Protect your assets in the next market
crash with StormGuard-Armor — it sets the standard for
market crash protection!
• USA Diversified ETFs Strategy
True Sector Rotation Strategy using IVE, IVV,
IVW, IWB, IWV, RSP, SPY, SPYG, MDYG and MDY.
(TLH bear symbol)
|
|
|
|
• ETF SPDR Sectors Strategy
True Sector Rotation Strategy using XLE, XLF, XLK,
XLI, XLP, XLV, XLY, XRT, XHB, XPH, MDY and SPY.
(UST bear symbol)
|
|
|
|
• ETF World Regions Strategy
True Sector Rotation Strategy using DGT, EEM,
EFA, EPP, FEZ, IEV, ILF, IOO, MDD, MDY, SPY and QQQ.
(TYD bear symbol)
|
|
|
|
|
|
