vaccine pessimism?
or
selling "vaccine optimism" news?
see below Max Pain🧐
The last column is today's $SPX closing price
the OI data was from yesterday's trading
today's OI would be available tomorrow
$SPX correlation with various volume PCRs for different periods
10-day, 30-day to 200-day correlations
1 SPX vs SPX: corr always = 1 for all periods
2 SPX vs VIX: corr always negative for all periods
3 SPX vs Equity PCR: corr negative for 5 out of 6 periods
4 SPX vs Total PCR
Quiz: 😉
Correlation between:
SPX vs SPX OI-PCR = positive across all periods
from 10-day to 200-day correlation
what does this mean? 🧐
Tons of SPX puts implying/predicting bearishness or bullishness for $SPX approaching OpEx?
highest market-value / book-value ratio
= Price to Book ratio
= PB ratio
$TSLA PE (TTM) = 1062
and
Price-to-Book ratio = 32.8
top 6 price-to-book companies:
ADSK
CTXS
MAR
ZM
IDXX
DOCU
surprised? 🧐 ZM on the list.
Another shocking finding today:
top chart = $VIX vs Equity put/call ratio (blue)
bottom chart = $VIX vs $DIX (green)
we know both equity-PCR & $DIX are positively correlated with $VIX
VIX = 21.64 (Red) flat.
Equity-PCR = 0.38 (Blue)
DIX = 0.38 (Green)
amazing coincidence?🤣
How to trade Equity PCR?
Standardized (then normalized) Equity PCR Z-score
-1 sigma = Over-sold Zone
-2 sigma = Fear Zone
-3 sigma = Capitulation Zone
if PCR z-score reaching below -1.5 or -2 sigma
= high probability of a $VIX spike in days
scale-in VIX long lotto tickets🧐👌
For $SPX STFR lotto tickets:
above definition should be reversed.
-1 sigma = Over-Bought Zone
-2 sigma = Greed Zone
-3 sigma = Euphoria Zone
VIX stats from 2006: 3800 days
color coded "fear & greed" zones
Samples= 3800 days
mean = 19.40
stdev = 9.61
min = 9.14
25% = 13.14
50% = 16.37
75% = 22.54
max = 82.69
2020 $VIX high ~= 6.5 sigma
intraday VIX high = 7 sigma
green = greed
red = fear
IF $VIX daily return is a perfect normal distribution (options modeling assumption) then:
6 sigma = every 1.38 million yrs
7 sigma = every 1.07 billion yrs
we had two in 10 years, and three in 33 year
the options modeling vastly underestimated the left tail risk🦃
🦃 day Quiz:
speaking of the "left tail risk" or "turkey 1001-day", how to estimate the left tail risk from options chains?
KEY:
it is called "SKEW"
1 negative put $SKEW = fat left tail (hedging for crash)
2 positive call $SKEW = fat right tail (hedging for melt-up)
for a given expiry series:
negative $SKEW = OTM put premium (or IV) is much higher than the OTM call premium
meaning people buying more puts for down-side protection & willing to pay high premium, driving OTM put IV higher than OTM call IV
positive $SKEW = OTM call IV > Put IV
Since 1987 crash, all fund managers have been extremely cautious, using index options to hedge their long stock portfolio to avoid another crash (VIX ~= 150 est.)
hence,
since 1987, all index options (SPX, SPY, QQQ, IWM etc) exhibiting much higher $SKEW than stocks options
How to estimate IV $SKEW from a given expiry series?
from $SKEW, we can guess "bullishness/bearishness" of a stock/ETF
many methodologies using "IV smile" curve
1 by OTM put & call delta (+-25 delta)
2 by OTM moneyness (+-10% from ATM strike)
3 by put-IV slope & call-IV slope
my own methodology:
4 I use the combination of the above three methodology
5 average the three using the below formula for normalization
6 (OTM put IV - OTM call IV) / ATM call IV
7 by normalizing (divided by) ATM call IV, it's possible to compare $SKEWS among different assets
Now, real-world examples
$AAPL Jan 15 expiry
IV smile
$AAPL IV SKEW is almost zero= no fun
Options players are neutral
three IVs: 10% moneyness
A OTM call IV (10%) ==> right side of ATM stk
B OTM put IV (-10%) =-> left side of the ATM stk
C ATM call IV
SKEW = (A - B) / C
For $AAPL above, I was using +-10% moneyness to calculate OTM call IV & OTM put IV vs ATM call IV
$SPHB outperforming XLK, QQQ, SPY, SPYG, VLU, SPYV.. you name it.
What could go wrong?
Maximum Optimism++ 🦃
From above Z-score charts, the "value" & "low volatility" alpha factor investing is dead for now
the momentum factor investing has always been superior, even before Fed QE-4ever
many research papers manifesting "momentum factor" from 6-12 months would continue to out-perform
to prove the "momentum factor" investing
below table contains perf key metrics for various Momentum periods
1 momentum factor based on "WQSR" ranking
2 lookback periods from 30-day to 300-day momentum
3 QQQ-103 stocks from 2013 to 2020
4 rebalance weekly selecting top-20 stocks
In a 60-40 portf, 60% of assets are invested in stocks & 40% in bonds- often government bonds. The reason it has been popular is that traditionally, in a bear market, the bond portion has functioned as insurance by providing income to cushion stock losses
more reasons than that:
I have been able to simulate 30 years of SPY & Government bond historical data using the following instruments; very close.
simulated SPY_1X = VFINX from 1990 to 2020
simulated Bond_1X = VUSTX from 1990 to 2020
compare the portfolio allocation performance for the last 30 yrs
important concept for the portfolio managers and private investors in order to increase alpha and reduce risk.
Quantitative analysis with real performance numbers.
comparing different allocation % between the stock and bonds for maximum returns and minimum risk & drawdowns.
30 yrs:
for SPY_1X = 1x leverage for SPY
CAGR=10.05%
Max Drawdown=55.23%
Sharpe Ratio=0.669
for Bond_1X=1x leverage in Government Bond
CAGR=6.91%
Max Drawdown= 19.36%
Sharpe Ratio=0.672
Bond_1X has a higher Sharpe ratio and lower drawdown
but with lower CAGR as well
Here is the Excel table from last 30 years of historical data for SPY_1X & Bond_1X as separate portfolios:
not surprised at all
as expected.
higher return in SPY would also incur higher risk (volatility and drawdown)
then,
what is the best % allocation between the two?
max drawdown is the killer
most mutual funds can't allow more than 50% drawdown
all investors would exit the funds
then the question is
what is the best pct % allocation between SPY & Bond for 30 years?
past results would not guarantee the future performance
but useful
From 1990 to today
for a traditional 60-40 stock to bond allocation:
CAGR = 9.17%
Max Drawdown = 32.39%
Sharpe Ratio = 0.902
not bad. like magic
CAGR increased to 9.17%
Max Drawdown reduced to 32.39%
and most importantly:
Sharpe ratio = 0.902
higher than both SPY and Bond.
This is the magic of diversification
increase CAGR and reduce risk/drawdown
and maximize Sharpe-Ratio
These are the real numbers
from 1990-2000 bubble
from 2001-2002 Tech wreck
GFC in 2008
then QE-4ever the last 10 yrs
60-40 still works like a charm
Now, the fun part
what if I had a crystal ball in 1990 & knew the best pct% allocation between the two?
well in hindsight, with ML optimization algos, I was able to come up with the below "best" allocation:
35-65 SPY to Bond
CAGR=8.34%
Drawdown=16.19%
Sharpe=1.01👍
perfect
summary table for everything we've discussed in one place
In hindsight:
it is no brainer that we would chose the 35 to 65 allocation ratio:
60-40 is not bad either.
the KEY is the lower drawdown & higher Sharpe Ratio.
with minimum CAGR penalty.
sleep well at night. 👍🧐
NYSE total Short Interest (SI) as of Nov 30 reporting
NYSE total Short Interest (SI) vs $SPX
NYSE SI and SPX Correlation = -63%
super high inverse correlation.
at extremes, it became one of the best contrarian indicators
All shorts, including monkeys/dogs capitulated. 🐂
VIX vs CBOE Equity PCR ratio
positively correlated most of the time
the pattern is clear
when $VIX up, people buying more puts (high PCR) for hedging
$VIX dn, euphoria buying all calls
$VIX vs Dark Pool Short Interest.
positively correlated some of the times
I report you decide
food for thought: to algo traders😉
1 algo for BTD on high Sharpe R & low vol stocks
2 trend & volatility filters
a 1 yr Sharpe > 1
b > 200d SMA
c pass if 5-d stdev > 2 * 50-d stdev or
d pass if 5-d ATR > 2 * 50-d ATR
e avoid recent high vol stocks
3 if a b c/d ==True
4 ..
4 entry criteria
5 MACD histogram crossover 0 line &
6 today's SMA(30d) < SMA(30d) 20 days ago &
7 making sure, the stock is on short-term pullback
8 calculate 20-d ATR
9 enter long with trailing stop at 4 X ATR(20)
10 taking profit or loss when price < 4 x ATR(20)
11 good
👌🧐
filtering out the high volatility stocks and down-trending stocks.
below is just 1 example of using 200d SMA trend filter to avoid 2002, 2008 & mar 2020 free fall market for SPX swing trading.
For stocks, trend & volatility filters are very important.
1 Machine Learning for Factor Investing R Version
by Tony Guida; Sep 1, 2020
2 Algorithmic Trading: A Practitioner's Guide Paperback – by Jeffrey M Bacidore; July 20, 2020
3. Machine Learning for Algorithmic Trading: Predictive models to extract signals from market and alternative data for systematic trading strategies with Python, 2nd Edition Paperback – July 31, 2020
by Stefan Jansen
- this is the best one for 2020 from my reading list.
For newbie Algo traders. Free
a great introductory algo book written by a real trader & well-respected hedge fund mgr, who divulged some of his most profitable strategies
Trading Evolved: Anyone can Build Killer Trading Strategies in Python Kindle (free)
why "volatility always came in clusters"? people ask
50% fundies & 50% technicals:
1 fundies/shocking-news triggering the first wave spike
2 people not prepared for the shock, scramble to buy protections
3 sell risky assets
4 rotate to safe-havens
6 which in term trigger the over-leveraged weak-hand longs, hitting stop loss levels & margin calls- 3d wave
7 risk-parity funds portf rebalance due to sudden plunge in risky assets. trimming
#2 of 3
8 60+% of commodity/futures leveraged funds, employing momentum/trend based strategies w/ trend detection & stop-loss levels for flipping to short, got flipped, huge force to the dn side: 5th wave