Business Handbook - Tortoise Capital



“Bounce” Trading System – An RLCO-based System for Low Volatility Markets

by

Kim Andersson

System Name: Bounce.

System Objectives: SQN (normalized to 100 trades) of at least 3; average R multiples of 2R/day, 5R/week, and 20R/month; generates 10 – 20 trades per day with holding periods of 5-15 minutes; functions in all longer-term market types (especially well in quiet market types), but may not perform as well in volatile sideways markets.

Position Sizing Objectives: Make at least 100% per year with a 5% chance of having a maximum 10% peak to trough drawdown during one month. What is important to me is consistency in returns. I want this system to be my daily bread and butter, workhorse type of system that consistently turns at least a weekly and preferably daily profit. This isn’t a system that will often hit home runs, but it should easily reach first base most of the time.

System Description: Bounce is a discretionary rules-based day trading system that uses Dr. Ken Long’s Regression Line Crossover (RLCO) “lens” to view the market and frame trades. The objective is to be positioned to take advantage of a price signal in my favor if I can frame the trade with a favorable reward:risk (R:R) ratio. In a nutshell, it takes advantage of “bounces” off of the Bollinger Bands, as well as the 30 period regression lines to capture small moves in the direction of the trend, which is determined by the slope of the 30 period +/-1 standard deviation Bollinger Bands. It is essentially a micro trend following system that takes advantage of pullbacks to and bounces off of bands. See the following screenshot for some examples of bounces off of the 30-period +/-1 SD BB and BB mean on a 5-minute chart.

[pic]

SQN: The SQN for a series of 197 live trades performed from the beginning of January 2015 until March 2015 resulted in an SQN (normalized to 100 trades) of 4.6. See the attached Excel spreadsheet for a list of the R-multiples, system expectancy, R standard deviation, SQN, and an R-multiple histogram.

[pic]

System Type: Rules-based discretionary day trading system consisting of a combination of band and micro trend following system.

Market Types: This system should work in all market types, but it is best suited for quiet markets and will probably not perform as well in volatile sideways markets.

System Edges: Frequent trading opportunities, low risk ideas, ability to cut losses short (sometimes less than 0.5R), a greater ability to trade in flow with the markets than a pure mechanical system because I am looking at price actions in several different periods on one chart, low correlation to my other systems, which are designed to work best in more volatile market conditions.

Key Beliefs:

• Regression lines and Bollinger Bands combine the insights of moving averages, the discipline of descriptive statistics, and a framework to pinpoint important critical states in different timeframes.

• Large ETFs (like SPY, DIA, and TLT) have a tendency to “bounce” off of key indicators such as Bollinger Bands.

• Large ETFs (like SPY, DIA, and TLT) are more likely to revert to the mean than individual stocks.

• Large ETFs (like SPY DIA, and TLT) are less “trendy” than individual stocks, but that does not mean that they don’t trend.

• Large ETFs (like SPY, DIA, and TLT) often move within relatively tight bands and are less volatile than individual stocks.

• This is not a mechanical system. I must be able to stalk, trade, monitor, and take profits (or abort trades) for one or more ETFs during the day session (9:30 AM to 4:00 PM ET).

• This system will likely provide at least a dozen trading opportunities per day depending on the timeframe used.

• Trading experience and market intuition are key factors in the successful execution of this strategy. In other words “Trader Quality Number (TQN)” (as opposed to the System Quality Number or SQN) is extremely important to the success of this system.

• I must always remember that exits matter much more to the overall profitability of this system than entries.

• Keeping losses very small (i.e., no larger than -0.5R) is very important in improving the SQN of this system.

• I don’t need to see price moving in my direction before I commit; I only look for a reasonably sized pullback which gives me enough evidence that a good reward:risk ratio exists.

• I want the best return in the shortest time period and hope for a 2:1 reward to risk ratio in less than 30 minutes.

• I assume that institutional money will come into the trade at perceived points of value (support) during the day.

• This system can go either long or short; however, I will (generally) only take trades that are in the direction of the overall trend as measured by the 5 minute charts and the 30 period Bollinger Bands.

• I believe I can improve my edge by being able to recognize a few chart patterns that recur regularly throughout the day.

• This is an intraday strategy. I expect to be in a position measured in terms of minutes from entry to exit. Trades that last 5 – 15 minutes are common. I want to move my stops to “no lose” (or breakeven) within the first 1 - 5 minutes, if possible. Otherwise, my belief is that the “bounce” I was anticipating is not going to happen, and I am better off preserving my capital for another opportunity.

• If the trade is not going to work, it’s not unusual for my initial stop to be hit within 1 to 5 minutes, thus taking me out of the trade and giving me a -0.5R loss. If the trade doesn’t move up to at least +1R or hit my initial stop in the first 30 minutes, then I should (probably) exit the trade. By cutting my losses short (and not waiting for them to hit full -1R losses), I can improve my results as measured by my R-multiple distribution, expectancy, SQN, and TQN.

• Context is important. The recent trend in the ETF (daily and hourly) are useful in identifying the direction that I should be trading in and what I can reasonably expect in terms of price action. For example, if SPY is in a bull market and breaks to a new high, it may find little resistance to continuing the rally to higher highs on an intraday basis. With a strong trending, quiet bull market, +5R to +10R returns are possible intraday.

System logic, concepts, and definitions:

• This system only trades large ETFs (like SPY, DIA, and TLT). SPY is the ETF that represents the S&P500 index; DIA is the ETF that represents the Dow 30; TLT is the ETF that represents the 20-year Treasury bond.

• “Short term” means intraday moves that are typically less than 60 minutes and often less than 30 minutes. I generally use a 1-minute (1”) intraday chart with candlesticks showing the price action. I use the 3-minute and 5-minute charts to gauge support and resistance levels, as well as the short-term trend.

• Entry – see below for when and how to enter trades.

• Initial Stop (istop) is the initial amount I put at risk and is the price between my entry price and my initial stop or exit price. Typically, it is $0.10 – $0.20 below (or above) my entry price.

• Range Stat: Ken uses the concept of range stat in his Frog system. It represents a price move of 5 Frog Standard Deviations, which Ken believes is a reasonable maximum expected intraday move for a large ETF or Dow stock.

• Reward:Risk Ratio is calculated as:

(Target Price – Entry Price) / (Entry Price – Initial Stop Price)

• Profit Targets:

o The first profit target is the 30-period Bollinger Band mean.

o The second profit target is the +/- 2 Standard Deviation of the 30-period Bollinger Band.

o The recent highs (or recent lows) on a 3 or 5-minute chart or other logical points of resistance or support (e.g., resistance level from 60 minute or daily chart).

o The VWAP (Volume Weighted Average Price).

• I want the best return in the shortest time period and would love to see a 2 to 1 return in 15 minutes or less.

Trading population:

• SPY, DIA, and TLT due to my belief that these are more likely to want to “bounce” off of key areas of support and resistance than individual stocks.

Setup rules:

For the setups, I want to see price at a key support or resistance level such as the following:

• Bollinger Band Bounce: price is approaching the +/-1 or +/-3 standard deviation Bollinger Bands. I also want to generally only take trades that are in the direction of the slope of the Bollinger Bands. The only exception to this rule is after a strong move, such as a full range stat move or more.

• RLCO (Regression Line Crossover) 10/30 Bounce: The following must occur for this set-up to be valid:

o The 10 period regression line crosses the 30 period regression line outside of the river (i.e., +/-1SD 30-period Bollinger Band) and is moving towards or into the “river”.

• HOD/LOD (high of day/low of day) support/resistance: Look for support/resistance at previous HOD or LOD.

• Determine if the trade meets the 2:1 Reward to Risk ratio. Use the Initial Stop, Entry, and Profit Target to calculate the Reward:Risk ratio. If it provides a minimum of 2:1, then I can frame the trade.

Entry rules:

If the setup is valid and Reward:Risk is greater than or equal to 2:1, then enter when:

• For Bollinger Band bounces: two entries are possible as follows:

o Enter as price pulls back or up to the +/-1 standard deviation Bollinger Band (SD BB); enter in the direction of the trend in accordance with the 5-minute chart and the slope of the BB.

o Enter as price touches the +/-3 SD BB on the 5-minute chart; enter in the direction of the BB mean.

• For HOD/LOD bounces: Enter when price touches the HOD/LOD or previous day’s HOD/LOD; enter in the direction of the BB mean.

Exit rules:

• Market stops you out:

o Initial Stop is hit

o Trailing Stop is hit

• Trailing Stops:

o Move the stop to “no lose” once I have +0.5R in hand. Now, I can monitor/manage the trade knowing I have nothing to lose.

o Move the stop up in 0.5R increments.

Optional decision rules:

• For more aggressive trades, I can choose to trade IWM (the ETF that represents the Russell 2000 small cap companies). This should only be done on very low volatility days, when IWM is outperforming SPY.

Position sizing rules:

• Example for a $0.1 stop: $250 at risk, $0.10 from entry to stop

o $250 / ($0.10/share) = 2,500 shares

• I will start trading this system with no more than 0.05% of my total account equity at risk. After 100 trades and a good R-multiple distribution, I will move my risk level per position to 0.1%. I will continue to move up my risk level in increments of 0.1% after each series of 100 trades with decent R-multiples, until I reach a maximum risk level of 1% of my total account equity per position.

Position sizing analysis results from Know Your System (KYS):

Below are the optimizer results from running 500 trades (maximum number of trades that KYS can simulate).

 

[pic]

 

So at 1% risk per position, I would have high confidence of seeing a return greater than 200% on just 500 trades with a negligible chance of hitting a 10% drawdown. Therefore, it’s likely that at 1,000 trades and 0.5% risk, I could meet my objectives with even lower drawdowns.   

 

The following figure represents the drawdown results in terms of R.

 

[pic]

 

The following figure represents the probable number of consecutive losses.

 

[pic]

 

 

Scaling-in rules: No scaling in with this system.

Trade management:

• It is quite likely that I will have several trade opportunities each day. It’s not unreasonable to look forward to as many as 50 or more trade opportunities in a week. Given that, I should not worry about a well-executed trade where I followed the rules and took a -0.5R loss – another trade opportunity is right around the corner. The losses are just the cost of doing business. In fact, losses in this system can be a very good indication of a change in trend; therefore, they are very useful tools.

• When I am in the trade, experience and intuition will go a long way to improving my profits through trade “efficiency”, or getting the most out of a move in my direction, while cutting losses short.

• When my trade moves close to my profit target, I am never wrong to use a “Zeno Stop.” With a Zeno stop, I evaluate the reward to risk ratio by looking at my profit target relative to the current price of my position, instead of relative to my current stop. For example, if price has moved in my favor, and I am currently looking at a 2R gain with a target that is 0.1R away from the current price, why risk 1R (from current price to my current stop) to get the remaining 0.1R out of a trade from that point forward? The reward:risk is not worth it, so I can close the trade and take the profit. The Zeno Stop is named after Zeno's (arrow) Paradox ().

Preferred brokers:

Low cost transaction per share, low slippage, and high speed of execution are the most important criteria for these systems. Large deep discount brokers should satisfy these requirements.

How I plan to start the system from all cash:

I will paper trade the system for three months (or at least 100 trades) until I am comfortable with the number and types of signals I receive, and that I can respond to the signals and set the buy/sell orders and trailing stops. Then, I will trade at reduced position size with real money at a deep discount broker in order to build confidence and professionalism with real money before increasing the position size incrementally and gradually.

Additional beliefs supporting this strategy:

• Markets trend and tradable trends exist on intraday charts.

• SPY, DIA, and TLT tend to act relatively predictably at key support and resistance areas such as Bollinger bands and HOD/LOD.

• SPY, DIA, and TLT do not like to move in extreme ways. They often seek to return to the mean.

• It is possible to have a profitable system trading large ETFs that have relatively low volatility.

• The RLCO lens works on multiple time frames, e.g., from 1 minute to daily charts.

• Prior points of support and resistance are useful signposts for identifying when price may turn around and go higher (at support) or stall (at resistance).

• The system can be profitable if I can frame the risk:reward as 2:1 or better.

• I take on substantially more risk if I hold positions overnight, so I exit positions before the end of the day session.

• I can gain an edge by trading only a few instruments that I get to know in-depth and understand how they behave.

How Will I know the System is Broken? If the SQN falls below 2, the system may be broken.

What Could Go Wrong with this System?

• This system is highly dependent on my skills as a trader as well as my psychology, so if I’m not operating at near 100% efficiency when trading this system, it may not work.

• This system will likely underperform in sideways volatile markets.

Chart examples:

Bollinger Band bounce chart example (i.e., entries at +/-1 SD BB and BB mean):

[pic]

RLCO 10/30 bounce chart example:

[pic]

HOD bounce chart example (i.e., entries at the daily HOD):

[pic]

LOD bounce chart example (i.e., entries at the daily LOD):

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download