Vertical Credit Spread Back Testing Study

1. Introduction
2. Summary
3. Rules
   1. Chart Studies
      1. Weighting the Studies
      2. Example A
      3. Example B
      4. Example C
      5. Example D
      6. Example E
      7. Example F
      8. Example G
      9. Example H
   2. Qualifying Securities
   3. Risk Management
   4. Routines
      1. Daily
      2. Monthly
   5. Entry Rules
      1. Bullish Entry
      2. Bearish Entry
   6. Exit Rules
      1. Bullish Exit
      2. Bearish Exit
4. Backtest Protocols and Results
   1. Backtest from May 2003 through present:
   2. Observations from Test Process
5. Original Set of Trades and Summary
6. Analysis of Losing Trades
   1. AMGN
      1. 2004/04/15
      2. 2004/11/12 – Not a bounce
      3. 2007/07/12
      4. 2008/01/08 – Counter Trend
      5. 2008/06/06
      6. 2008/10/21
   2. DIA
      1. 2005/02/22 – Mis-recorded – was small winner.
      2. 2005/04/06
   3. GENZ
      1. 2003/10/24
      2. 2004/02/03
      3. 2004/03/31
   4. QCOM
      1. 2003/09/18
      2. 2003/11/07
      3. 2003/12/09 – Counter Trend, No Signal
      4. 2004/05/26
      5. 2004/12/22 – Stop Ignored
      6. 2005/10/27
      7. 2006/02/21
      8. 2006/09/07
      9. 2006/10/11
      10. 2008/07/15 – Trade near earnings
      11. 2008/12/01
   5. WFT
      1. 2004/03/01 – Past Resistance
      2. 2004/05/18
      3. 2004/06/24
      4. 2004/12/14 – Weak Entry
      5. 2005/03/01 – Counter Trend
      6. 2005/03/03 – Stop Ignored
      7. 2005/10/12 – Past Support, Ignored Entry
      8. 2005/10/14
      9. 2005/10/20 – Past Support, Ignored Entry
      10. 2005/12/19 – Counter Trend, Ignored Entry
      11. 2008/02/15
      12. 2008/03/24 – Mis noted, closed normally
7. Corrected Trades and Summary

Introduction

In March 2009 we formalized our rules for Credit Spread trading and executed an extensive back test of the rules.  The back test involved running over 200 trades from mid 2003 to Mar 2009 on 5 securities.  The testing also helped to refine the rules to some extent and identify weaknesses.

Directional trades are generally more profitable, but if the trend is sideways then directional trades are problematic.  Provided there is enough volatility in the underlying security, credit spreads are a good way to profit from price movements that directional trades may not profit as easily with. 

Credit Spreads are long of Theta and short of Volatility.  They are profitable if the price moves in the direction favoring the trade, or if the price remains flat but keeps the short option out of the money.

Revisions

7/2009 - Clarifyied uses of technical studies.  Some are used more for screening than for trade decisions and should not be part of the trade decision.

4/2009 - Updated examples to provide guidance on the more discretionary aspects of the system.

3/2009 - Initial revision documents the baseline rules, studies, and the backtest results.

Summary

After correcting some errors and correcting for some broken rules the population of 200 trades yielded only 20% losing trades, 45% of trades executed without modification, and there was an average return on risk of 11.6% on an average trade of 37.9 days, annualized this is around 111.7% return on risk.  The results with all the errors were 8.8% gain for 37.4 days or about 85.9% annualized.

Either way this looks like an acceptable strategy for some portion of a portfolio.  Trading the DIA was easier than a single stock - the fund is less volatile and dollar strikes make risk management easier.

Rules

Rules help remove emotion, but there is still judgment on occasion.  These rules include discretionary exits which can limit gains though they are intended to mitigate losses.

We show the study set we use and a handful of entry/exit points.  These are the rules we back-tested.  The live rules running on the Strategies page may have changed...

Chart Studies

Below is the study set we generally use for this strategy.  We have 1% and 2% above the high and below the low included as helpers, the 10, 30, 50, and 200 SMA, and an SMA envelope shown on the 30-DMA with a 3% offset, but we'll commonly change that, using different averages as needed adjusting to trends on particular securities.

You will also see an EMA Bollinger Band light in the background and also the $SPX index light in the background.  We use candles and the averages more than these, but have them for reference.

We use the Investools standard MACD and Stochastic studies, but also use Slow Stochastic (80,20) which has served us well.

We also use Accumulation/Distribution, On Balance Volume, and Wilder RSI as supporting indicators, but rely more strongly on candles and averages - everything else is just supporting evidence.

Volume is present because it can be used to help verify the formation of price patterns which take multiple days to form.

This is a chart of ABT from Q1 of 2009.  The stock has been channeling sideways between about $48 and $58 for most of the last 2 years.  The October 2008 minimum is right around $45 and the horizontal price lines have an interval of $2.5 for reference.

In the chart I have identified some potential entry/exit signals based on price breakouts of support/resistance:

1. Bullish Entry on breakout above moving averages.  MACD and Stochastic upturn, after earnings.
   
2. Potential Bearish Entry, crosses below the 10-DMA.  The MACD histogram has been falling off, and the Signal begins a downtrend the next day.
3. Bearish Entry with Support Breakdown, or a chance to pile on more.  MACD makes sharp turn confirming the change.
4. Possible Bullish Entry on support bounce with candle pattern (Piercing Line).
   
5. Potential Bearish Entry (and trend confirmation) bouncing down around the 30-DMA.  The SS line confirms this before the MACD study.
6. Potential Bearish Entry, or a chance to pile on more.
7. Bearish Entry with Support Breakdown, or a chance to pile on more.
8. Potential Bearish Entry, or a chance to pile on more.

Any Bearish Entry is also a potential place to take profits and/or close a Bullish Trade.  For example the event (D) may be a place to take profit in trades from the first 2 events, while the event (E) may be a place to exit the trade from (D) since it suggests a continued downtrend.

Falling off the end of the chart we may, however, have pretty strong support forming.  This may be cause to keep a close leash on any bearish trades and not enter new trades until there is a breakout.  Alternatively this may be an area to have trades on for both directions - ready to close either on a breakout of the flat price pattern.

Weighting the Studies

• Price is KING - Flags are First - we like them supported by confirmed candles like those shown in the charts above.
• We view every bounce (flag breakout) or price pattern breakout or support/resistance breakout as a potential entry/exit point.
• If the momentum change is supported by a reversal candle pattern, so much the better to help confirm the trade.
• Without candles we'll go with entry 1% past the high/low of the low/high day.
  
• If we see a potential entry/exit point we look to potential support and resistance, weighing each pretty equally:
• Horizontal and Sloped lines which may reach back several months.
• The moving averages.
  
• The formation of the current price pattern itself, perhaps slightly stronger as they are recent.
  
• It is best if several of these are converging to support the direction of the new trade.
• If we seem to have a bounce at support/resistance then we look to the MACD/Stochastic studies for confirmation:
• We like to see 2 of the 3 technical indicators supporting the direction of the trade.
• When we have a potential reversal of the intermediate trend we look to the AD/OBV/RSI
• These help to indicate if a security is being accumulated or is losing momentum.

We try to avoid running a trade through earnings as these are short-term trades.  However, in back testing we could only see the hard right edge of the chart and couldn't look ahead to earnings.  This did not contribute to many of the significant losing trades.  We only really examined trades that lost more than 10% of the max loss, so this might have contributed to more of the smaller losses.

Our back testing generally went with entry of trades at the end of the day.  Instead of entering trades at pre-determined thresholds, we would set alerts, and those alerts would give us the opportunity to enter new trades at the end of the day if the price held - or enter or exit in the middle of a day with a huge price movement.

Example A

The stock has bee channeling sideways for a long time now, we tend to look for entry of trades around the ends of the price range.

This is a Moving Average Breakout following earnings.  The problem here is that it breaks into what is likely to be a strong and broad band of resistance made of of the highs from October and November, and frankly for the last 2 years.

This has areas of confused support and resistance and is probably worth avoiding a trade which is in the middle of the price range.

Example B

The anticipated resistance we found in (A) seems confirmed.  This is a place for a potential bearish entry with the price stalling around an expected top of the price range.

This is supported by the MACD histogram, but not so much by the Stochastic studies; the STO(14,5) is slightly weakening, though.

RSI shows a pause, but AD and OBV have been rising and it is unclear if they are flattening or not.

All in all we are at the top end of the price range with a sideways channeling overall price pattern.  This makes a new bearish trade reasonable.

Example C

If you entered a trade at (B), this confirms the trade.  If you were waiting for a break of resistance, support is broken instead.

It looks like resistance held, and the 10-DMA is now swinging down.

By this time the MACD lines are down trending, and the signal has bounced off the average line (histogram never went positive).  The Stochastic line looked like it might bounce, but turned down, and the Slow Stochastic signal never even got near the average line.

Also, by this time, the AD/OBV/RSI studies have gone flat, RSI is even slightly down.  This suggests the breakdown may be real since accumulation has stopped.

Even if you entered a trade at (B), this confirmation may be worth adding a position to the trade.

Example D

This is a potential bullish entry, particularly in the context of the October 2008 low, which these 2 candles pierce and rebound from.  The candle pattern is a Piercing Line.  Buying volume is lower than the selling volume on the previous day, though.  The 30-DMA is now swinging downward, and the 50-DMA even shows a slight down movement.

The MACD lines are about to cross, but the smoothed line is low and just flattening.  The Stochastic study is oversold, and the Slow Stochastis signal line may be about to hook up.

AD and OBV have leveled at a higher low than the previous.  RSI is a little weaker.

A piercing line at support with the MACD about to turn up and following a relative pause in selling.  This is at least a signal to consider taking profits from (B) and (C).

It may be early to enter a bullish trade, but in our back-testing we would have.

Example E

If you entered a bullish trade at (D), it may be turning against you.  If you still have trades from (B) and (C), you were right.

We have bounced off the January low as resistance.  The 30-DMA (actually about 3% below it) is serving as resistance.  The candle pattern is a confirmed Bearhis Harami.

The MACD Histogram has weakened, the the Slow Stochastic signal turned sharply to hug the smoothed line.

AD/OBV/RSI never strengthened from the signal at C, suggesting this price swing was a flag and not a reversal.

Probably time to take profit from the bullish trade on C and potentially time to enter a new bearish trade.  While it is around the middle of the trading range, there is space between the moving averages and the October low, so it can stay in this lower channel for a while.

One point against this trade is that we are starting to get close to earnings.

Example F

Example G

Downtrending resistance held and the stock broke below the October 2008 lows.  The price pattern is a failed new high confirming a bear flag followed by a break of Support.  There is no strong candle pattern to cite here, but the flag pattern is broken.  The 30-DMA is confirmed as resistance for now.

The difference between this and (D) is that the previous candle closed at resistance, and this candle opened at resistance, tried to run up and broke back down.  In (D) the previous day's candle did close below the drawn resistance line, but the oscillators aren't supporting a bearish trade.

The MACD here supports the downturn, as do the Stochastic studies.

AD/OBV/RSI never strengthened from the signal at C, and have dropped some more.

If you didn't enter a bearish trade at E, this looks like a good entry, or a place to pile on with more.

The problem is that earnings is just a few days away now

Example H

Now earnings are over, price action confirms a resistance bounce, and if you didn't enter a trade at (G) then this may be a good entry, though there are signs of support around $42 now.

The Stochastic oscillators have remained low and diverted from an upturn, supporting a bearish trade.  The histogram isn't signaling much at all, and this condition rarely lasts very long.

There's no real sign of accumulation suggesting a support bounce is too likely.

Qualifying Securities

Stock or ETF that are range bound or reversing are best.  For stocks an Implied Volatility of at least 35% seems to be needed.  Backtesting of DIA suggested that only about 16% Implied Volatility was needed to find the desired range in premiums for risk.

If the stock is strongly trending then apply a more directional strategy.

It is here that we also use the Accumulation/Distribution, On Balance Volume, and Relative Strength technical indicators.  These are used with initial screening and with re-screening for a new trade.  These indicators can be used to help identify changes to the trend of a security.

????????????????????????????????????????

Risk Management

Backtesting was with trades having a maximum risk of $1000 to $1500, typically 3 contracts.  We prefer not to risk more than 2-3% of the total portfolio or 5-10% of the specific account.

We are looking for a gain of no less than $0.66, typically $1 to $2 on a strike spread of $5 - Minimum Max Gain of 15%, typically 25% to 67%.

Resistance prevents upward price movement. The short strike of a Bear Call Spread should be at or above a price representing resistance.
The Bear Call Spread is profitable provided the price remains below the short strike price. Price movement can be flat or downward.
	This illustrates the guidelines for selection of the short strike price, defining NTM/ATM above. This is QQQQ from Feb to Mar 2009. ETF's often have dollar strikes making it easy to get close to the candle.  In this chart the grid is on dollar increments. Most stocks have 5-dollar strikes, so a strike overlapping the candle may be the best compromise.
Price movement can be flat or upward. The Bull Put Spread is profitable provided the price remains above the short strike price.
The short strike of a Bull Put Spread should be at or below a price representing support. Support prevents downward price movement.

Routines

Daily

Evaluate trade daily, reconsider normal exit or adjustment points.

Also consider discretionary exits; if there is no compelling reason to exit, let it run, compelling reasons include:

• Have >50% of profit with <50% of time - may be becoming more volatile than normal.
• Have >80% of profit and may reverse - may as well capture profit and move on.
• Consider moving exit threshold up/down on subsequent flags.

Set up orders or alerts for the following day.  Backtesting generally suggested use of alerts - triggers would often be tripped at bad exit points, but waiting often produced a better exit point.

Monthly

Evaluate the recent population of trades (say last 3-6 months).  From the backtest population of 200 trades we have the following expectations:

• The average gain was 11.6% with a standard deviation of 17.9%
  
• About 20% losing trades and 80% winning trades
  
• About 55% of trades exit with some sort of modification, 45% by covering the short option for a nickel
• Across each security AND with all compined trades in the study there were a maximum of 3 consecutive losing trades.
  

Allowing 2 Standard Deviations the average gain may be from (6.3%) to 29.5% (Average +- 2 Standard Deviations).  The larger the population the less volatile the result will be.

Smaller populations will be more volatile and more likely to deviate from the normal population.

Strategy rules, or use under current market conditions, should be re-examined if any of the following conditions arise:

• Win/Loss or Normal/Modified exits change by more than about 5%
• Return on Risk changes by more than about 10%
• There are more than 3 consecutive losses
  

Note that correlation to the market may result in losses at the same time, this may suggest a need for risk management rule adjustments but probably not a failure of the strategy.

Entry Rules

We view every bounce (flag breakout) or price pattern breakout or support/resistance breakout as a potential trade entry point.  It is permitted to have multiple trades at the same time, but testing suggests the occurance of signals to do this is relatively rare on single stocks and somewhat more likely on index ETFs.

Consider a risk control switch called TSAR (Trade Support And Resistance).  When TSAR is OFF we only take trades in the direction of an established trend.  When TSAR is ON we will consider trades at both support and resistance bounces because there is no clear trend, or the trend may be reversing, or at the limits of an observed price channel.

The TSAR switch is turned OFF when a trend can be identified, then we enter trades on bounces or breakouts supporting the current trend.  Confirm a trend by:

• When the 30 and 50 DMA lines tend to be parallel in their slope.
• When the Stochastic and MACD smoothed lines tend to stay flat but are above or below the centerline.
  

The following conditions turn TSAR ON and we enter trades on both Support and Resistance bounces - preferably at the limits of an observed price channel:

• A reversing price pattern such as failed new high/low, or a Double Top/Bottom or Head and Shoulders.  Confirmation of these patterns can be supported with volume, too.
• A horizontal or only slightly sloped price channel in comparison to the top-to-bottom range of the price swings.
  

If the breakout is not confused then we look to the Oscillator studies MACD/Stochastic/SS and we want to have 2 signals supporting the trade.

If the breakout is counter trend or suggesting a trend change we will look to the Accumulation Momentum studies AD/OBV/RSI for additional support suggesting the trend is potentially changing.  When these lines are flat they indicate no significant change in the flow of big money.  Upward slopes suggest money going in, and downward slopes suggest money going out.

Note that a Bounce at Support/Resistance is often the playing out of a Flag at the Support/Resistance level and often indicated by a candle reversal pattern.

Note that a Breakout is often playing out a "Flag in a stiff breeze" prior to the breakout.

Finally, we traditionally avoid entering trades close to earnings.  This wasn't specifically part of the back test process, and was not a factor in most of the larger losing trades we investigated closely, but may be a contributor to some of the trades which lost less than 10% of the max loss, or required modified exits with less than the max gain.  What is not speculation is that many stocks have volatile price swings around earnings and Implied Volatility inflates prior to earnings and usually deflates rapidly afterwards.  On the surface this sounds ideal for credit spreads, but you hafe to get the direction right...

Select strikes, bearing in mind the guidelines in Risk Management.

Bullish Entry

If the breakout is with an upward price movement off of support or through resistance, with no obvious resistance close by, then the entry is supported and we look to the oscillator studies.

We want to have 2 supporting the trade:

• MACD - If the smoothed line is low and flat and the signal line is crossing up, or the smoothed line is up trending, or the smoothed line is high and flat.  When the smoothed line is high it may be late in a trend, but the line will stay high if there is a strong trend running and we use the histogram (line crossovers) for signals.
  
• Stochastic(14,5) - This study only has the smoothed line.  If the line is low and turning up, or the line is up trending, or the line is high and flat.
  
• SlowStochastic(80,20) - If the smoothed line is low and flat and the signal line is crossing up, or the smoothed line is up trending, or the smoothed line is high and flat.  When the smoothed line is high it may be late in a trend, but the line will stay high if there is a strong trend running and we use the histogram (line crossovers) for signals.

If the trade is a counter-trend trade in a potential reversal we also look to the AD/OBV/RSI studies for support of the trade.  Upward movement in these signals suggests big money moving in.  If the stock has been trending down, but these signals have been flat for a while that downward price movement may be losing the support of big money.

Bearish Entry

If the breakout is with a downward price movement off of resistance or through support, with no obvious support close by, then entry is supported and we look to the oscillator studies:

We want to have 2 supporting the trade:

• MACD - If the smoothed line is high and flat and the signal line is crossing down, or the smoothed line is up down trending, or the smoothed line is low and flat.  When the smoothed line is low it may be late in a trend, but the line will stay low if there is a strong trend running and we use the histogram (line crossovers) for signals.
  
• Stochastic(14,5) - This study only has the smoothed line.  If the line is high and turning down, or the line is down trending, or the line is low and flat.
  
• SlowStochastic(80,20) - If the smoothed line is high and flat and the signal line is crossing down, or the smoothed line is down trending, or the smoothed line is low and flat.  When the smoothed line is low it may be late in a trend, but the line will stay low if there is a strong trend running and we use the histogram (line crossovers) for signals.

If the trade is a counter-trend trade in a potential reversal we also look to the AD/OBV/RSI studies for support of the trade.  Downward movement in these signals suggests big money moving out.  If the stock has been trending up, but these signals have been flat for a while that upward price movement may be losing the support of big money.

Exit Rules

We view every bounce (flag breakout) or price pattern breakout or support/resistance breakout as a potential trade exit point.  Ideally, though, we want the trade to run out over time allowing us to gain most of the maximum possible gain.

Consider the risk control switch called LIR (Let It Ride).  If we are in a strong trend we tend to let trades run to a normal exit, permitting flags to move against the trade because they are short and small.  However, if price action has been choppy and/or there has been a losing trade or two then consider LIR to be off, and we exit at any potential exit point.

The LIR switch is turned ON when a trend can be identified, then we allow trades to run to normal exit (buy back short option for $0.05).  Confirm a trend by:

• When the 30 and 50 DMA lines tend to be parallel in their slope.
• When the Stochastic and MACD smoothed lines tend to stay flat but are above or below the centerline.
  

The following conditions turn LIR OFF and we take profits when the price action shows weakness for the trade:

• A reversing price pattern such as failed new high/low, or a Double Top/Bottom or Head and Shoulders.  Confirmation of these patterns can be supported with volume, too.
• The previous trade was a losing trade.
• The previous trade would have been a losing trade if it had not been exited early on signs of weakness.

In the back test study we often took profits early.  We developed the risk control described above to define the conditions where we tended to have a modified exit.  In general we used the end of day prices for the exit conditions, but on days when there was a strong breakout we made the assumption that we would at least have alerts to notify us of conditions where the trade was showing weakness; the alert would allow us to decide how to manage unexpected price action.  In most of those conditions we assumed we would have set an order based on the underlying price if the price continued to move strongly against the trade.  When this happened we would use the Think Back tool to estimate the intra-day option price.

Using programmed orders, though, we would often have exited (and entered, for that matter) trades when it was not necessary.  This is why we chose end-of-day trading, it was more than just making the back test process quicker and easier.  For example, you have entered a Bear Call spread, but the stock gaps up dramatically in the morning.  The price action then drops for the entire day, ultimately forming a Dark Cloud or Bearish Engulfing candle pattern.  An automated order would probably have cut you out at a loss, but an alert would allow you to observe and decide - if the price gapped up and ran away you could choose to exit as the price action continued against the trade.

Had we employed programmed stop orders there would have been several more large losing trades, and the study shows how just a few trades can significantly affect the profitability of the strategy.

If not using a stop order, then set an alert at a price to allow evaluating an exit order if momentum moves suddenly against the trade. 

If using a stop order then make it an OCO order with the other order.  Back testing was generally done covering the short option at the stop and holding the long option until the first sign of weakness against the trade.

Instead of stop or outright exit, consider adjustment, typically by applying a butterfly to roll the strike up/down.

Example of trades gone bad.  XME from Dec 2007 into Jan 2008.  The stock is in an uptrend from a low in August 2007 which retraced 50% of the run in Nov 2008.  These are not necessarily losing trades due to risk management strategies in the entry and exit rules... Entry 1 is a support bounce off the moving averages at a price level that was resistance in Oct 2007. A possible Double Top forms, though.  This causes us to be more conservative and take profits when the price starts to fall off.  The short option from (1) is covered and a new trade (2) is entered in the opposite direction. The Double Top confirms and there is another support bounce at the 200DMA which is also a previous support level from November, as well as being around the price target for the Double Top.  Here we sell the long put from (1), cover the short call from (2), and enter a new trade (3). This fails quickly, however, and we sell the long call from (2) and cover the short put from (3), but this is not a good entry since it is so close to likely support. A few days later a hammer forms in a dark candle with an upper shadow.  The following day is a light candle hammer, but in a lower price range.  The day after that is a third hammer at almost the same price level with no upper shadow.  It seems clear this is forming support, and volatility is deflating, so we sell the long put from (3).  This is still not a clear entry signal because the hammer candle pattern is not confirmed. On the next day the hammer is confirmed and we can enter a new trade (4) which ultimately exits normally...

Bullish Exit

???????????????????????????????????

Bearish Exit

???????????????????????????????????

Backtest Protocols and Results

Follow rules above. Without access to risk profile, changed them to describe max risk instead of using risk profile.
It's hard to get concurrence with industry group from past, so just going with stock trades.

• Test from Mar 2003 to present.
• Set chart to 10Y/Daily.
• Enter the stock symbol.
• Quickly scan the chart for no more than about 5 seconds looking for enough reversals and range trading to make a number of spread trades possible.
• Keeping the price chart in peripheral vision, look only at the dates and zoom in to the 3-6 months previous to Mar 2003.
• Set the zoom to about 3 months for comfortable viewing of candles and recent trends, begin looking for an entry point.
• It's OK to zoom out to see back in time, AVOID SEEING ANY FUTURE DATA.
• Advance a day at a time until you find entry/exit conditions and record price data.
• If triggered entry/exit would likely be used then interpolate intra-day price, otherwise it's quicker to use end of day.

Use the MARK/THEO price - the formulas are set to fill at spread mark-0.05 - where we typically get filled...
Closing trade use the ASK
We aren't recording sell of long option - usually worthless and we'll do better in reality if we sell any for anything at all.

Note:  This last point handicaps the results to some extent, but a few trades did account for the long option sale offsetting losses where it made sense.

Backtest from May 2003 through present:

• AMGN - Steve
• QCOM - Steve
• GENZ – Kimber (2003-2005) poor IV made this hard to trade
• WFT – Kimber w/ Steve
• DIA – Steve (2004-2005) filled out population

Observations from Test Process

This is pure technical trading – we have no fundamental analysis from the past.  While these stocks have been on our IBD based watch lists in the last 12 months we do not know how they would have scored out in the past.

Kimber requires help / more practice identifying flags and reversals and generally interpreting candles and technical indicators.  She will start with Bulkowski’s self-test regimen at thepatternsite.com.

Single Stocks require Implied Volatility > 35%

DIA required Implied Volatility > 16% -- DIA was easier to trade and generally less volatile.

Original Set of Trades and Summary

Following is the table summarizing the trades generated in the back test.  The first set are the losing trades.  The second set are those trades which had a modified exit but resulted in profit.  The last set are the trades that just went perfectly right.

Analysis of Losing Trades

15 of the trades (winners and losers) do not meet the 15% minimum max return on risk.

It is recognized that some of the winning trades may not have been under ideal conditions, but our analysis will focus on the trades in the back-test which lost more than 10% of the maximum risk in the trade.

Most of the trades are just trades gone bad – it happens.

Statistically speaking there are likely as many errors among the profitable trades.

AMGN

2004/04/15

This Bull Put trade was entered on a break above the 30-DMA.  The stock was pulling off a support level in what appeared to be a channeling pattern.  The price bounced down off the 50-DMA a few days later and the trade was exited.  MACD and Stochastic indicators support the entry.

2004/11/12 – Not a bounce

This Bear Call trade was entered on exit of a Bull Put trade in or near resistance.  This resistance held for 13 days and failed with a price breakout.  This trade was not entered on a confirmed bounce from resistance.

2007/07/12

This Bull Put trade was entered on a bounce (above the high of the low day) off of a higher low.  The trade was near potential resistance from the moving averages.  The trend remained flat until 2 days after the earnings release.  The stock price had been stable in spite of strong downward pressure in the S&P 500.

The subsequent opposite trade exited early with gains far less than the maximum on another reversal.

2008/01/08 – Counter Trend

This Bull Put trade was a counter-trend trade entered on a bounce from a slightly lower low after a strong down-run.  The bounce did not result in a trend change and exited with a bounce off resistance at the 30-DMA.

2008/06/06

This Bear Call trade was entered on a bounce (below the low of the high day) near a horizontal and a descending line of resistance.  Because of the very narrow price channel an exit at the support bounce a couple days later might have been prudent.  At this stage the moving averages were beginning to turn up.  The MACD and Stochastic upturns were also early warnings for a better discretionary exit.

2008/10/21

This Bear Call trade is entered on an unconfirmed shooting star.  The pattern did confirm the next day, but hit a strong reversal.  The trade was adjusted with a butterfly.  In retrospect covering the short call and subsequently selling the long call would have resulted in a lesser loss.

DIA

2005/02/22 – Mis-recorded – was small winner.

This Bear Call trade is entered on a break through moving averages while the stock was in a flat range.  The short call was covered 2 days later as the breakdown failed.  Several days later the stock bounced from a higher high and a new call was sold, this trade completed normally.

2005/04/06

This Bull Put trade is entered on a support bounce, horizontal support and 200-DMA.  The long put was sold several days later, mitigating the loss.

Because of nearby support an opposite trade was not entered when this trade went bad.  A new trade was entered at the end of the month which completed normally.

GENZ

2003/10/24

This Bull Put trade is entered on a dragonfly doji (unconfirmed) supporting a same low bounce.  The trade is exited at the original stop, but it is clear the trade is going bad with a confirmed resistance bounce on 10/30 off the 50, 30, and 10 DMA.  This is a tough call nearby support suggests possibly adjusting down if the price was right.

The subsequent trade from a 200 DMA bounce ended normally.

The possible missed trade on 10/17 would have to play the NOV 50, paying $2 this was a little too close to the money.

2004/02/03

This Bull Put trade is a 10-DMA support bounce which looks to break to new highs.  The trend failed to continue, but support seemed to hold until the breakdown where the trade was closed.

2004/03/31

This Bull Put trade is possibly counter-trend, but entered on a horizontal support bounce.  The stock appears to be trading in a range.

The resistance bounce on 4/8 should have been a signal the trade was going wrong.  This was a resistance bounce off the 200 and 30 DMA and breakdown through the 10DMA.  The other technicals (particularly SS) suggest the trade is counter-trend and failing.

QCOM

2003/09/18

This Bull Put trade was entered after a break above long-term resistance with an uptrend and volume on the breakout day.  I don’t see a way to better this – the trade turned bad quickly, but the stop was a break of the resistance used as an entry trigger.

2003/11/07

This Bull Put trade was entered after a small flag bounce with an uptrend.  The stock held flat and the volume fell off.  The large bearish candle on 11/18 was the first to break the stop.

2003/12/09 – Counter Trend, No Signal

This Bear Call trade was entered counter-trend at a long-term horizontal resistance level without a good candle confirmation of a resistance bounce.

2004/05/26

This Bear Call trade was entered near recent resistance in a flat trading range, but without a bounce signal.  The price did pause for a few days, then continued up.  A better entry would have been around 6/18 when it rolled down from a higher price.

An opposite trade was entered on 6/22 with a strong support bounce.

It is worth noting this trade also counted sale of the long call, which most all of the trades do not…

A good bull put trade was missed on 7/22

2004/12/22 – Stop Ignored

This Bull Put trade was entered at resistance on a moving average support bounce with a volume spike.  The trade should have been stopped earlier 12/30 instead of letting it run along the stop level…

2005/10/27

This Bear Call trade was entered near a long resistance line, on a break down through moving averages, and a failure to make a new high.  There was a gap down followed by support, in 2 trading days made over 50% profit from short call.  With a natural tendency for a gap to fill, probably should have taken profit.  It would have been better to exit when the gap filled than to ride.  These are all discretionary exits, though…

2006/02/21

This Bear Call trade was entered near a long resistance line on a break down through the 30 DMA and a failure to make a new high. 

2006/09/07

This Bear Call trade was entered after a higher high with a higher low, the 30-DMA had turned up.  This did look like a possible resistance bounce from several months ago.

It was, perhaps, an iffy trade given the moving averages, but I’m not sure anything is wrong.

2006/10/11

Mismarked, exit was at .60, not 1.25 – 1.0% loss.

2008/07/15 – Trade near earnings

This Bear Call was entered after a lower high.  There was a large volume up day before after-hour earnings.  Could have exited on that clue with profit and eliminated the earnings risk.

2008/12/01

This Bear Call trade was entered on a trend continuation bounce off the 30-DMA.  Selling the long call would have offset some of the loss, but this looks like a simple losing trade.

WFT

2004/03/01 – Past Resistance

This Bull Put trade was entered on a continuation bounce, but around previous horizontal resistance.  It failed at the resistance bounce.

2004/05/18

This Bear Call trade was entered on a moving average bounce, but at a price which was both recent and past resistance, and sitting on the 200-DMA.  It failed at the support bounce.

2004/06/24

This Bear Call trade was entered on a resistance bounce around where the 3/1 trade failed.  The candle confirms a hanging man at resistance.  STO and MACD are overbought.  Setup seems fair.  The breakdown failed and was followed by a bullish breakout.  Might have exited a day earlier on discretion.

2004/12/14 – Weak Entry

This Bull Put trade was entered on a break above the Moving Averages, possibly a couple days late given a recent trading range, but good to confirm not a resistance bounce – it’s still not a great breakout.  It failed on the break back below the averages.

2005/03/01 – Counter Trend

This Bear Call trade was entered off a new high in an uptrend.  Actually, compared to previous runs, it is pretty old so the timing may look good.  The stock bounce on the up trending 10DMA leading to the next trade.  This was a risky entry, but a quick exit…

2005/03/03 – Stop Ignored

This Bull Put trade was entered off a continuation bounce in an uptrend.  As mentioned in the previous trade the run was a bit aged, but the previous entry was probably early.  The stop for this trade was ignored.

2005/10/12 – Past Support, Ignored Entry

This Bear Call trade was entered on an apparent continuation pattern, but the day’s candle ends right at old support, and there was just a down move roughly equal to recent price fluctuations.

The better entry was on or before 10/5 where the stock rolled off highs and broke through the moving averages support, however this is a counter-trend trade, too.

2005/10/14

This Bull Put trade was entered on a support bounce on the 200DMA and around previous support.  There is a resistance bounce at the 10DMA which undercuts the stop and breaks the 200DMA.

2005/10/20 – Past Support, Ignored Entry

This Bear Call trade is entered on the break of the 200 DMA, but is still around strong support.  Selling the long call on 11/7 would have significantly offset this loss.

Probably due to frustration over the last 3 trades, the correct Bull Put trade is ignored when this one fails.

2005/12/19 – Counter Trend, Ignored Entry

This Bear Call trade is entered on a break of the 10-DMA, but is counter trend and the 30DMA has been strong support lately.

The better entry is the Bull Put on the day this trade went bad.

2008/02/15

This Bear Call trade is entered on a resistance bounce and a lower/similar high, but the bounce fails and ends in a breakout.

2008/03/24 – Mis noted, closed normally

This Bull Put trade exits normally on 4/21.

Corrected Trades and Summary

Most of the losing trades above are just trades gone bad.  Some have notes in the title indicating a broken rule.  In the table below losing trades which violated the rules are modified or removed, also trades that were ignored originally were added to the table.  By simply eliminating the worst rule breakers the change is dramatic:

Charts - All Charts are screen captures with permission of Prophet Charts.