C Crucil Kansas City Research Weekend June 17 2016 Background Trading for three years intraday swing and long term Trading channeling system for past 15 years What I have learned about myself as a trader ID: 591299
Download Presentation The PPT/PDF document "Mechanical trading of channelling swing ..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Mechanical trading of channelling swing system
C. Crucil - Kansas City Research Weekend – June 17, 2016Slide2
Background
Trading for three years – intraday, swing and long term
Trading channeling system for past 1.5 years
What I have learned about myself as a trader:
My results show no evidence that my intuition/discretion provides an edge so currently choosing to use mechanical rules that I can
backtest
so that I can isolate source of problems as system, market or self
Finding a trading system that fits the individual is imperative - I am currently focusing on swing and long-term systems which suits my strengths (analytical observer) and weaknesses (slow decision maker)
I have not achieved consistent success in trading but do have some areas of moderate success – I am choosing to build upon these potential beachheads
Although I realize that losses are an inevitable part of trading, I have discovered a low tolerance for drawdown – I am a
brickmaker
not a homerun hitterSlide3
Rolling average nSQN
of past 100 CH trades
Things working well as a starting point; relatively small drawdowns were alleviated by either different market conditions or improved stock screening.
!?!?!?!Slide4
Baseline Channelling Backtest
Current standard rules:
Enter long
when Williams %R < 20 (oversold) and close > 200 Day MA (long-term
trend); enter on open next day following signal
Exit on open next day when Williams %R > 70
Backtest only considers one entryUniverse: ETF30, S&P100 and NASDAQ100 (155 total)Test Period: January 1, 2011 – March 31, 2016
Results:# Trades: 6,139Expectancy: 0.16R Win Rate: 59% nSQN: 1.16
Total Return: 1,104%Max Drawdown: 82%Slide5
Considerations for rule adjustment
Primary objective is to:
reduce max drawdown (82% in baseline) and
to reduce portfolio risk on days when high number of signals fire (i.e. fewer high quality entries)
Options for rule adjustment:
Overbought/oversold indicator choice (i.e. RSI1, NDX or Williams %R)
Screen for intermediate price trend rather than just long-term (i.e. 200D MA)Chart pattern more likely for successMethods to reduce open risk with multiple consecutive correlated trades (system generates many signals with over 50% of the stock universe in a single day)
Role of short/intermediate market direction (system currently only long in bull market)Role of market volatilitySlide6
Shorter-Term Trend
Considering we are looking for a mover over a 2-10 day period, would it be beneficial to consider a shorter term time frame for the definition of an uptrend?
Uptrend criteria options:
Close > 200
DayMA
(baseline)Close > 100 Day MA
Close > 50 DayMASlide7
Short/Intermediate Trend Results
# Trades
Expectancy
nSQN
(100)
Win Rate
Total Return
Max Drawdown
Baseline (200 Day MA)
6,139
0.16R
1.16
59%
1,014%
82%
Baseline + 100
Day MA
4,3050.17R1.2660%734%51%Baseline + 50 Day MA3,1600.16R1.2161%491%56%Slide8
Overbought/oversold Indicators
Williams %
R (10):
enter next day at open when
%R
< 20 and close > 200 Day MA ;
exit next day at open when %R > 70RSI2: enter next day at open when RSI2 < 20 and close > 200 Day MA e
xit next day at open when RSI2 > 70NDX:enter next day at open when NDX < 20 and close > 200 Day MA ;
exit next day at open when NDX > 70
Baseline
William %r < 20 +
close > 200
DayMA
Plus 100
DayMA
Plus 50
DayMA
New BaselineWilliam %r < 20 + close > 200 DayMA + 100 DayMARSI2NDXSlide9
Overbought/oversold indicator results
# Trades
Expectancy
nSQN
(100)
Win Rate
Total Return
Max Drawdown
Baseline (Williams %rR)
6,139
0.16R
1.16
59%
1,014%
82%
New Baseline (Williams %
rR
)4,3050.17R1.2660%734%51%New Baseline + NDX2,0200.18R1.4357%361%31%New Baseline
+ RSI2
1,726
0.21R
1.49
60%
364%
24%Slide10
Chart Patterns
Two chart patterns appear in 20% of all signals: 1) big red bar and 2) gap down
Signal after a gap down
Signal after big red barSlide11
Chart pattern Results
# Trades
Expectancy
nSQN
(100)
Win Rate
Total Return
Max Drawdown
New Baseline + Big red
candle
318
0.35R
2.42
63%
New Baseline
+
Gap
870.05R0.3163%Slide12
Still too many entry signals most days
At 1% position sizing, ~ 60% of portfolio at risk with open positionsSlide13
Stock Selection Screens
Objective is to generate a maximum of 3 correlated entries per day
Stock selection screens:
Reward:risk
ratio: reward =10 day high; risk = 1 ATR (14)
Relative volatility: ATR (14) % relative to SPY
Relative strength: 30 day change relative to SPY Type of symbol (i.e.Dow30, Nasdaq, S&P, ETF)
New BaselineWilliam %r < 20 + close > 200 DayMA +
close > 100 DayMA + RSI2 < 20Excluding gaps
Reward: Risk
Relative Volatility
Relative Strength
Type of symbol Slide14
Stock Screen
F
ilter results
Reward: Risk
Relative Volatility
Relative Strength
Symbol
category
<2
2-3
>3
<0
0-1
>1
<0
0-5
5-10
>10Dow30NasdaqS&PETF# Trades
82
808
750
89
984
516
679
607
265
89
nSQN
1.15
0.95
2.25
1.48
0.86
2.73
1.38
1.70
1.26
2.91
0.66
2.1
1.26
2.33
Avg.
0.14R
0.13R
0.32R
0.15R
0.1R
0.46R
0.19R
0.22R
0.20R
0.55R
0.08R
0.35R
0.17R
0.22RSlide15
Market Direction
Basic rule set uses 200D MA and market classification system enter longs only in bull market
Considering we are looking for a mover over a 2-10 day period, would it be beneficial to consider a shorter term time frame for the definition of an uptrend
?
Market direction screens:
Market classification system (bull, bear, sideways)
SPY 100 day moving averageSPY slope of 30 period linear regression line
New BaselineWilliam %r < 20 close > 200 DayMA
close > 100 DayMA RSI2 < 20
Excluding gaps
Max three entries per day based on R:R,
RelVol
&
RelStr
Market Classification
SPY 100
DayMA
SPY 200 DayMASPY slope of 30 regression lineSlide16
Market Direction Filter results
Market Classification
SPY 100 D MA
SPY
Slope30
Bull
Bear
Sideways
Above
Below
Above 10d
Avg
Below 10
d
avg
# Trades
381
4568352141109385nSQN 2.263.514.722.373.73
2.58
2.81
Avg.
0.34R
0.70R
0.85R
0.36R
0.69R
0.35R
0.47RSlide17
Market Volatility
Market volatility screens:
Market classification system (quiet, normal, volatile)
RiskZ
VIXSlide18
Market Direction Filter results
Market Classification
RiskZ
Quiet
Normal
Volatile
>0.50
0.50 – (0.50)
<
(0.50)
# Trades
75
281
91
130
179
185
nSQN 3.971.974.572.972.113.25Avg.0.52R0.31R
0.9R
0.42R
0.34R
0.56RSlide19
Revised Ruleset
Enter long at next open based on following criteria:
close > 100 day MA
RSI2 < 5
Williams %
rR< 20Exit next open when Williams %R > 70Exclude entries after larger gaps
Max 5 entries per day (sort by Rel Vol and R:R)No entries when market volatility = normal & RiskZ > (0.05) and < (0.50)Slide20
Equity Curve
CAGR = 25%Slide21
Return/Drawdown
2011
2012
2013
2014
2015
Number of Trades
88
112
67
77
56
Return
50.8%
18.35%
24.25%
20.9%
14.1%Max Drawdown13.0%12.6%2.1%3.2%4.8%Expectancy0.61R0.28R0.68R0.64R0.74R
nSQN
3.84
1.86
6.48
3.29
4.31Slide22
Summary
System is more sensitive to market volatility than market direction; system performs better at volatility/
RiskZ
extremes
System requires extensive entry screening to reduce portfolio risk
Study does not consider exits;
63% of losses are full 1R lossesReducing drawdown comes at the cost of lower return (?)