Testing the "Buy in November, Sell in May" Strategy

Test Parameters and Assumptions

Strong Seasonality Trade (November-May) With Filters

Strong Seasonality Trade (November-May) With Filters

Utilizing the MACD filter and comparing it to our baseline system, we can see we don't make as much more as the simple moving average filter, however, we increased our risk adjusted return with a NP vs DD of 3.8. So, if drawdown is your concern and not just pure profit it looks the MACD filter is slightly better.

RSI Filter

For our final filter I will try the RSI indicator with its default loopback period of 14. Again, like the MACD calculation, I want price moving in our direction so I want the RSI calculation to be above zero when opening a position and below zero when closing a position.

Strong Seasonality Trade (November-May) With Filters

Weak Seasonality Trade (May-November) With Filters

The inefficient way:

if c > c[1] and c[1] > c[2] and c[2] > c[3] then
buy somewhere tomorrow

value1 = patternTests - 1;

if(value1 > 0) then
begin

if(mod(value1,2) = 1) then patternBitChanger[0] = 1;
value2 = value1 - patternBitChanger[0] * 1;

if(value2 >= 8) then begin
patternBitChanger[3] = 1;
value2 = value2 - 8;
end;

if(value2 >= 4) then begin
patternBitChanger[2] = 1;
value2 = value2 - 4;
end;
if(value2 = 2) then patternBitChanger[1] = 1;
end;

Do you remember binary based systems where 0 – off and 1 – on? If you don’t remember its OK. Here is a quick review for everyone.

0 0 0 0 pattern = 0
0 0 0 1 pattern = 1
0 0 1 0 pattern = 2
0 1 0 0 pattern = 4
1 0 0 0 pattern = 8

Each 1 or 0 is a placeholder and its placeholder is a bit than can either be on or off. The leftmost bit is 1’s place. The next bit is the 2’s place. The next bit is the 4’s place. And finally the rightmost bit is the 8’s place. So if we plug in 1 1 1 1 we get (8 + 4 + 2 + 1) = 15. If we start counting at 0 [ 0 0 0 0 ] then we can represent 16 distinct patterns with our four bits. Following?

So using this bit scheme and TradeStation’s optimizer we can run 16 different patterns sequentially. Now how do we get the bit pattern scheme to relate to the last four day’s close to close relationships. This is where the eloquence continues [if I have to say so myself]. EasyLanguage has a library of powerful string functions. One of those is the ability to concatenate or add a string to a string.

Let’s pretend we are the computer and we are optimizing thru 16 different runs. Let’s call the first run [0 run]. In our tests 1’s will be represented by “+” and 0’s by “-“. Using our 4 bit scheme how to we represent the number 0? Zero means no bits are on so 0 = [0 0 0 0]. If we translate our bits to a string where 1’s are “+” and 0’s are “-” then we get “- – – -“]. Right? If a “+” represents an up close and “-” represents a down close then the string “- – – -” would indicate four straight down closes.

Run 0 = 0 0 0 0 = ” – – – – ” – four straight closes down
Run 1 = 0 0 0 1 = ” – – – + ” – three straight closes down followed by 1 up close
Run 2 = 0 0 1 0 = ” – – + – ” – two straight closes down, one up close, one down close
Run 3 = 0 0 1 1 = ” – – + + ” – two straight closes down, two straight closes up
…
…
…
Run 15= 1 1 1 1 = ” + + + + ” – four straight closes up

Following this scheme we can easily check out 16 different 4 day patterns. Let’s see what pattern is the most prolific on the long side. What do you guess – maybe after 4 consecutive down closes?

Simply buying the next bar’s open and exiting two days later, let’s see which pattern is king. Here is the pattern and its profit/loss:

Well it seems they were mostly winners but you shouldn’t buy after ” + + + + “ or "+ + + -". Pattern ” – – – +” was the best followed by ” + – – – “.

Oh yeah this was tested in the ES over the past five years so we are probably witnessing a bullish bias. Let’s see if we sell using the same patterns.

Well the bullish bias is definitely evident. The best 4 day pattern to sell the next bar is to wait for 4 consecutive up closes or 3 consecutive up closes followed by 1 down close.

Is there anything to this pattern recognition? I think it can be used to filter trades and that’s about it. We tested a four day pattern for illustration purposes only but that might be too many. A two day pattern might work better. Please check this out and let me know.

Here is the program in its entirety:

input: patternTests(14),orbAmount(0.20),LorS(1),holdDays(0),atrAvgLen(10);

var: patternTest(""),patternString(""),tempString("");
var: iCnt(0),jCnt(0);
array: patternBitChanger[4](0);

{written by George Pruitt -- copyright 2006 by George Pruitt
This will test a 4 day pattern based on the open to close
relationship. A plus represents a close greater than its
open, whereas a minus represents a close less than its open.
The default pattern is set to pattern 14 +++- (1110 binary).
You can optimize the different patterns by optimizing the
patternTests input from 1 to 16 and the orbAmount from .01 to
whatever you like. Same goes for the hold days, but in this
case you optimize start at zero. The LorS input can be
optimized from 1 to 2 with 1 being buy and 2 being sellshort.}

patternString = "";
patternTest = "";

patternBitChanger[0] = 0;
patternBitChanger[1] = 0;
patternBitChanger[2] = 0;
patternBitChanger[3] = 0;

value1 = patternTests - 1;

if(value1 > 0) then
begin

if(mod(value1,2) = 1) then patternBitChanger[0] = 1;
value2 = value1 - patternBitChanger[0] * 1;

if(value2 >= 8) then begin
patternBitChanger[3] = 1;
value2 = value2 - 8;
end;

if(value2 >= 4) then begin
patternBitChanger[2] = 1;
value2 = value2 - 4;
end;
if(value2 = 2) then patternBitChanger[1] = 1;
end;

patternString = "";

for iCnt = 3 downto 0 begin
if(patternBitChanger[iCnt] = 1) then
begin
patternTest = patternTest + "+";
end
else
begin
patternTest = patternTest + "-";
end;
end;

for iCnt = 3 downto 0
begin
if(close[iCnt]> close[iCnt+1]) then
begin
patternString = patternString + "+";
end
else
begin
patternString = patternString + "-";

end;
end;

if(barNumber = 1) then print(elDateToString(date)," pattern ",patternTest," ",patternTests-1);
if(patternString = patternTest) then begin

// print(date," ",patternString," ",patternTest); //uncomment this and you can print out the pattern
// if(LorS = 2) then SellShort("PatternSell") next bar at open of tomorrow - avgTrueRange(atrAvgLen) * orbAmount stop;
// if(LorS = 1) then buy("PatternBuy") next bar at open of tomorrow + avgTrueRange(atrAvgLen) * orbAmount stop;

if(LorS = 2) then SellShort("PatternSell") next bar at open;
if(LorS = 1) then buy("PatternBuy") next bar at open;

end;

if(holdDays = 0 ) then setExitonClose;
if(holdDays > 0) then
begin
if(barsSinceEntry = holdDays and LorS = 2) then BuyToCover("xbarLExit") next bar at open;
if(barsSinceEntry = holdDays and LorS = 1) then Sell("xbarSExit") next bar at open;
end;

-- George Pruitt from GeorgePruitt.com.

If you have been involved with the markets for even a short period of time, you have heard about this trade.  Buy N days prior to the end of the month and then exit M days after the end of the month.  This is a simple test to perform if you have a way to determine the N and the M in the algorithm.  You could always buy on the 24th of the month, but the 24th of the month may not equal N days prior to the end of the month. 

Simple approach that doesn’t always work – buy the 24th of the month and exit the 5th of the following month.

If you have been involved with the markets for even a short period of time, you have heard about this trade. Buy N days prior to the end of the month and then exit M days after the end of the month. This is a simple test to perform if you have a way to determine the N and the M in the algorithm. You could always buy on the 24th of the month, but the 24th of the month may not equal N days prior to the end of the month.

Simple approach that doesn’t always work – buy the 24th of the month and exit the 5th of the following month.

if dayOfMonth(d) = 24 then buy next bar at open;

if marketPosition = 1 and dayOfMonth(d) = 5 then sell next bar at open;

Before we get into a little better coding of this algorithm, let’s see the numbers. The first graph is trading one contract of the ES futures once a month – no execution fees were applied. The same goes for the US bond futures chart that follows. Before reading further please read this.

CFTC-required risk disclosure for hypothetical results:

Hypothetical performance results have many inherent limitations, some of which are described below. No representation is being made that any account will or is likely to achieve profits or losses similar to those shown. in fact, there are frequently sharp differences between hypothetical performance results and the actual results subsequently achieved by any particular trading program.

One of the limitations of hypothetical performance results is that they are generally prepared with the benefit of hindsight. In addition, hypothetical trading does not involve financial risk, and no hypothetical trading record can completely account for the impact of financial risk in actual trading. For example, the ability to withstand losses or to adhere to a particular trading program in spite of trading losses are material points which can also adversely affect actual trading results. There are numerous other factors related to the markets in general or to the implementation of any specific trading program which cannot be fully accounted for in the preparation of hypothetical performance results and all of which can adversely affect actual trading results.

No Pain No Gain
Looking into the maw of draw down and seeing the jagged and long teeth.

Draw down as a percentage of account value.

Detto!

The bonds had more frequent draw down but not so deep. These teeth can cause a lot of pain.

Well, George what is the N and M?
I should have done M days before and N days after to maintain alphabetic order, but here you go.

ES: N = 6 AND M = 6

US: N =10 AND M = 1

How did you do this?
Some testing platforms have built-in seasonality tools, but, and I could be wrong I didn’t find what I needed in the TradeStation function library. So, I built my own.

A TradingDaysLeftInMonth function had to be created. This function is a broad swipe at attempting to determining this value. It’s not very smart because it doesn’t take HOLIDAYS into consideration. But for a quick analysis it is fine. How does one design such a function? First off, what do we know to help provide information that might be useful? We know how many days are in each month (again this function isn’t smart enough to take into consideration leap years) and we know what day of the week each trading day belongs to. We have this function DayOfWeek(Date) already in EasyLanguage. And we know the DayOfMonth(Date) (built-in too!) With these three tidbits of information, we should be able to come up with a useful function. Not to mention a little programming knowledge. I was working on a Python project when I was thinking of this function, so I decided to prototype it there. No worries, the algorithm can be easily translated to EasyLanguage. And yes, I could have used my concept of a Sandbox to prototype in EasyLanguage as well. Remember a sandbox is a playground where you can quickly test a snippet of code. Using the ONCE keyword, you can quickly throw some generic EasyLanguage together sans trade directives and mate it to a chart and get to the nuts and bolts quickly. I personally like having an indicator and a strategy sandbox. Here is a generic snippet of code where we assume the day of month is the 16th and it is a Monday ( 2 – 1 for Sunday thru 7 for Saturday) and there are 31 days in whatever month.

currentDayOfWeek = 2;
currentDayOfMonth = 16;
loopDOW = currentDayOfWeek;
daysInMonth = 31
#create the calender for the remaining month
tdToEOM=0; #total days to EOM
for j in range(currentDayOfMonth,daysInMonth+1):
if loopDOW != 1 and loopDOW != 7:
tdToEOM +=1;
print(j," ",loopDOW," ",tdToEOM)
loopDOW +=1;
if loopDOW > 7: loopDOW = 1; #start back on Monday

Create a synthetic calendar from the current day of month

I just absolutely love the simplicity of Python.  When I am prototyping for EasyLanguage, I put a semicolon at the end of each line.  Python doesn’t care.  Here is the output from this snippet of code.

Cur>Day DOWDay DaysLeftAccum.
-----------------------------------
16        2    1 Monday
17        3    2 Tuesday
18        4    3 Wednesday
19        5    4 Thursday
20        6    5 Friday
21        7    5 Saturday
22        1    5 Sunday
23        2    6 Monday
24        3    7 Tuesday
25        4    8 Wednesday
26        5    9 Thursday
27        6    10 Friday
28        7    10 Saturday
28        1    10 Sunday
30        2    11 Monday
31        3    12 Tuesday
On Monday 16th there were 12 Trading Days Left In Month Inclusive

Output of Python Snippet - use in EZLang.

I start out with the current day of the month, 16 and loop through the rest of the days of the month. Whenever I encounter a Sunday (1) or a Saturday (7) I do not increment tdToEOM, else I do increment.

Here is how the function works on a chart. Remember in TradeStation I am placing a market order for the NEXT BAR.

Counting the days until the EOM

This snippet of code is the heart of the function, but you must make in generic for any day of any month. Here it is in EasyLanguage – you will see the similarity between the Python snippet and its corresponding EasyLanguage.

array: monthDays[12](0);

monthDays[1] = 31;
monthDays[2] = 28;
monthDays[3] = 31;
monthDays[4] = 30;
monthDays[5] = 31;
monthDays[6] = 30;
monthDays[7] = 31;
monthDays[8] = 31;
monthDays[9] = 30;
monthDays[10] = 31;
monthDays[11] = 30;
monthDays[12] = 31;

vars: curDayOfMonth(0),curDayOfWeek(0),loopDOW(0),tdToEOM(0),j(0);

curDayOfWeek = dayOfWeek(d);
curDayOfMonth = dayOfMonth(d);

{Python prototype
tdToEOM=0;
for j in range(currentDayOfMonth,daysInMonth+1):
if loopDOW != 1 and loopDOW != 7:
tdToEOM +=1;
print(j," ",loopDOW," ",tdToEOM)
loopDOW +=1;
if loopDOW > 7: loopDOW = 1;
}

loopDOW = curDayOfWeek+1;
tdToEOM=0;

for j = curDayOfMonth to monthDays[month(d)]
begin
if loopDOW <> 1 and loopDOW <> 7 then
tdToEOM +=1; // tdToEOM = tdToEOM + 1;
loopDOW +=1;
if loopDOW > 7 then loopDOW = 1;
end;
TradingDaysLeftInMonth = tdToEOM;

EasyLanguage Function : TradingDaysLeftInMonth

I used arrays to store the number of days in each month. You might find a better method. Once I get the day of the month and the day of the week I get to work. EasyLanguage uses a 0 for Sunday so to be compliant with the Python function I add a 1 to it. I then loop from the current day of month through monthDays[month(d)]. Remember month(d) returns the month number [1…12]. A perfect index into my array. That is all there is to it. The code is simple, but the concept requires a little thinking. Okay, now that we have the tools for data mining, let’s do some. I did this by creating the following strategy (the same strategy that create the original equity curves.)

inputs: numDaysBeforeEOM(8),numDaysAfterEOM(10),movingAvgLen(100);
inputs: stopLossAmount(1500),profitTargetAmount(4000);

vars: TDLM(0),TDIM(0);

TDLM = tradingDaysLeftInMonth;
TDIM = tradingDayOfMonth;

if c >= average(c,movingAvgLen) and TDLM = numDaysBeforeEOM then
begin
buy("Buy B4 EOM") next bar at open;
end;

if marketPosition = 1 and barsSinceEntry > 3 then
begin
if TDIM = numDaysAfterEOM then
begin
sell("Sell TDOM") next bar at open;
end;
end;
setStopLoss(stopLossAmount);
setProfitTarget(profitTargetAmount);

EasyLanguage function driver in form of Strategy

A complete strategy has trade management and an entry and an exit. In this case, I added an additional feature – a trend detector in the form of a longer-term moving average. Let’s see if we can improve the trading system. Thank goodness for Genetic Optimization. Here is the @ES market.

Get your Pick ready to mine!

Smoothed the equity curve – took the big draw down out.

Genetically MODIFIED – Data Mining at its best!

Here are the parameters:

Did not like the moving average. Wide stop and wide profit objective. Days to EOM and after EOM stayed the same.

Bond System:

Bond market results.

If you like this type of programming check out my books at Amazon.com. I have books on Python and of course EasyLanguage. I quickly assembled a YouTube video discussing this post here.

Conclusion – there is something here, no doubt. But it can be a risky proposition. It definitely could provide fodder for the basis of a more complete trading system.

>> By George Pruitt from blog georgepruitt.com

But this is one interpretation. Here is another way to think about it. Go long when the Money Flow rises above 80 and short when it falls below 20. See the image below.

However, these are not all possibilities. If you think about it, you can have various combinations of entry signals based on the upper and lower thresholds of the Money Flow Indicator. We could also introduce a midpoint value (50) that can act as a trigger.

For long trades, we could...

• buy entering the overbought region
• buy leaving the overbought region
• buy crossing above the midpoint (50)

I would classify these signals as momentum trades. The Money Flow is at the higher end of its range, and we're going long. 

On the other hand, we could fade these values. 

• buy entering the oversold region
• buy leaving the oversold region
• buy crossing below Midline (5)

How do we know which triggers to test? The answer is we don't. Thus we code up all of these into a switch statement using EasyLanguage. We then use TradeStation's optimization features to iterate over different buy/sell signal combinations to see which produces better performance.

The Money Flow Strategy

Below is the testing strategy to determine which buy/sell signals work best.

Input:
iBuyType(1),
iSellType(1),
iLookback(10);

vars:
MyMoneyFlow(0),
Oversold(20),
Midline(50),
Overbought(80);

MyMoneyFlow = MoneyFlow( iLookback );

Switch( iBuyType )
Begin
Case 1:
   if MyMoneyFlow Crosses above Overbought then buy next bar at market;
Case 2:
   if MyMoneyFlow Crosses below Overbought then buy next bar at market;
Case 3:
   if MyMoneyFlow Crosses above Midline then buy next bar at market;
Case 4:
   if MyMoneyFlow Crosses below OverSold then buy next bar at market;
Case 5:
   if MyMoneyFlow Crosses above OverSold then buy next bar at market;
Case 6:
   if MyMoneyFlow Crosses below Midline then buy next bar at market;
Default:
   Print("Invalid iBuyType Value");
End;

Switch( iSellType )
Begin
Case 1:
   if MyMoneyFlow Crosses above Overbought then Sellshort next bar at market;
Case 2:
   if MyMoneyFlow Crosses below Overbought then Sellshort next bar at market;
Case 3:
   if MyMoneyFlow Crosses above Midline then Sellshort next bar at market;
Case 4:
   if MyMoneyFlow Crosses below OverSold then Sellshort next bar at market;
Case 5:
   if MyMoneyFlow Crosses above OverSold then Sellshort next bar at market;
Case 6:
   if MyMoneyFlow Crosses below Midline then Sellshort next bar at market;
Default:
   Print("Invalid iSellType Value");
End;

The EasyLanguage code and TradeStation WorkSpace are available for you to download once you become an EasyLanguage Mastery Insider! It's free and you get access to all our code. See the bottom of this article for details.

Testing Environment

Before getting into the details of the results, let me say this: Unless otherwise stated, all the tests within this article are going to use the following assumptions:

• Starting account size of $100,000
• In-sample dates are from 2006 through December 2020
• Commissions & Slippage were not deducted
• One contract was traded per signal
• Overbought Threshold 80
• Oversold Threshold 20
• Midline Threshold 50

Testing Money Flow Heating Oil

I can now load my testing strategy onto a daily Heating Oil (@HO) chart and test it. I will optimize to input values. First, the Buy Type with values between 1-6. Next, the Sell Type with values 1-6. This will produce 36 different tests.

I organized the performance results by Perfect Profit Correlation, which measures the straightness of the equity curve. The best-performing settings for Heating Oil were Buy Type 6 and Sell Type 4.

Looking at the code for these input values...

• Go long when the Money Flow crosses below the Midline.
• Go short when the Money Flow crosses below the oversold region.

Here is an image of the Money Flow signals on the Heating Oil chart.

Below are the equity curve and performance report.

Nice! We found a potential starting point for a viable trading system. TradeStation recently released a micro contract for Heating Oil. Trading a daily system based around Money Flow may work.

Summary

In this article, I've explored the Money Flow Indicator, a technical analysis tool used in trading to measure the strength of buying or selling pressure in a security. Considering price and volume data, the Money Flow Indicator can help identify overbought and oversold conditions, providing traders with valuable insights into market sentiment. I've developed a testing strategy with various entry and exit triggers based on the Money Flow Indicator and tested it on the Heating Oil futures contract.

Conclusion

The testing strategy I've presented demonstrates that the Money Flow Indicator can effectively generate trading signals. By optimizing input values and examining various buy/sell signal combinations, traders can find the best-performing settings for their specific market. In the case of Heating Oil, going long when the Money Flow crosses below the Midline and going short when it crosses below the oversold region produced the best results.

Note this strategy is a testing strategy, not a complete trading system! The strategy is not for live trading but for helping you locate the best entry and exit triggers on the Money Flow Indicator. If you want to learn more about building trading systems, check out the article, How to Build Profitable Trading Systems.

To make this a complete strategy, you would need to add a stop value, test different filters, and then put the strategy through a validation process. 

Please join EasyLanguage Mastery Insider to download this article's EasyLanguage code and TradeStation WorkSpace.

Gap Zones as defined by Scott Andrews:

If the previous day was a down day:

• Zone 1 - Today's Price is above yesterdays high
• Zone 2 - Today's Price is between yesterdays open and high
• Zone 3 - Today's Price is between yesterdays open and close
• Zone 4 - Today's Price is between yesterdays close and low
• Zone 5- Today's Price is below yesterdays low

If the previous day was an up day:

• Zone 6 - Today's Price is above yesterdays high
• Zone 7 - Today's Price is between yesterdays high and open
• Zone 8 - Today's Price is between yesterdays open and close
• Zone 9 - Today's Price is between yesterdays open and low
• Zone 10 - Today's Price is below yesterdays low

Scott used this information to help filter gap trades for the current day. Depending upon today's Gap Zone, a given trade may tend to work out well or not. In some cases applying the Gap Zone as a filter dramatically proved trading performance. 

The question is, will it help us in the overnight session? 

Added The Gap Zone Code

First I'm going to create an input called GapZone which will be used to test each of the possible Gap Zones.

Input:GapZone(0);

If we set GapZone to zero, that will tell the strategy to ignore the Gap Zone filter. Otherwise, GapZone can hold the values 1-10 representing our ten Gap Zones.

Then, I'm going to create some variables to hold important price information.

I'm then going to use these variables to simply see what Gap Zone we're in based upon today's opening price. 

I'll also create a variable called TradeGapZone that will tell our strategy if today is a valid Gap Zone. For example, if we're testing Gap Zone 7 and today's price is Gap Zone 7 then TradeGapZone is set to true and trade is opened.

That's basically it.

The Over Night Edge on S&P Futures

Let's take a look at trading the overnight session on the @ES.D symbol. This test does not take into account slippage or commissions. There are no filters. We open a trade every night (15:15 PM Central) and close the market open (8:30 AM Central) the next day. Below is the equity curve.

We can now apply our Gap Zones filter. I will optimize the input parameter GapZone with the values 0-10 to test which Gap Zone(s) may produce the best results. Why start at zero? Remember, if we set GapZone to zero, that disables the Gap Zone filter. I do this so I can see the results without the filter alongside my other results.

Here is the optimization report organized in descending order based upon average profit per trade. Click the image for a larger view.

You can see that Gap Zone 5 is the best option. Only take trading with Gap Zone 5 produces an average profit of $155 per trade and the following equity curve.

A huge improvement!

What is Gap Zone 5? Looking at the Gap Zone chart above, Gap Zone 5 is when 1) The previous day was a down day and 2) Today's open was below yesterday's low.

That's an ugly trade, right? We have a down day yesterday and today's open is below yesterday's low. Yet, when you buy at the close and exit your position, you get the following results..

Adding A Trend Filter

As we enter the bear market of 2022, the overnight edge does not hold up either. Bull markets are where this edge is the strongest. How can we avoid unproductive bear markets? Adding our standard 200-day simple moving average as a regime filter is a filter known to work well. Let's add that.

This means we'll only take trades when the price is above a 200-period simple moving average. When the price is above the 200-period simple moving average, we are in a bull market and free to trade. When the price is below the 200-period simple moving average, we are in a bear market and don't trade. The equity curve and results are below.

We can see a considerable drawdown reduction and a solid increase in NP/DD, Profit Factor, and average profit per trade. Sure, we're making less money as we are trading less often. But overall, the improvements look good.

What Else Can We Do?

What other filters could we test?

Seasonality: You could create a simple filter that only takes trades during specific months. Thus, measuring the impact of different seasons on the overnight edge. Thinking along these lines, you could also test a day-of-the-week filter. Does Friday perform better than Thursday? Is Wednesday a day to avoid opening new trades?

Combine Zones: Looking at the optimization report once again, we can see the second choice is Zone 8. This zone happened when the previous day was an up-day, and today's price is between yesterday's open and close. You could combine Zone 8 and Zone 5. Thus, taking trades in either of these zones. This will add more trades and more profit.

Can We Turn This Into A Real Trading System?

Using Gap Zones can provide significant benefits to taking advantage of the overnight trades. Could this be turned into a real trading system? I think so!

I would move to a smaller timeframe, such as a 5, 10, or 15-minute chart. Moving to these smaller timeframes would allow me to enter trades after the close and better monitor the trades overnight. I would also explore what my stop would be. Another idea to test would be holding the trade until the next day's close vs. closing it at the open.

This should provide you with some great ideas to test on your own.

Best of luck!

Next, let's test the behavior above and below the regime filter. I performed an optimization for each of the situations utilizing the same range of values as above. Why split the market by a regime filter? My reasoning is market behavior changes between these two regime states. I picked the best value for each optimization based on perfect profit correlation. Here are the results

Bull Market
Profitable Tests: 114/155 (74%)
OpenDay = 3
HoldDays = 3

Bear market
Profitable Tests: 101/155 (65%)
OpenDay = 22
HoldDays = 3

Below is a table with all three results. 

When price is trading above the regime filter (Bull), we can see the first few days of the month hold the best promise. In this case, we're entering on the 3rd day and holding for three days.

However, when price is below the regime filter (Bear), more toward the end of the month is where we see our bias—in this case, entering on the 22nd day and holding for three days.

Below is the performance of each of these three different trading systems with their optimized values based upon perfect profit correlation.

These numbers look good. Looks like we have a promising trading system that is simple to trade. Anyone can do it!

You can imagine combing the Bull Only and Bear Only into a single killer strategy. I think we can build a good trading system from this study.

Is This Valid?

But before we do that, I would like to perform another test. My fear with our current analysis is all our data is in-sample, and this approach is not statistically valid.

Did you catch that? If you did CONGRATULATIONS!

Our results look great. It's tempting to add as stop and start trading. But that would be a mistake. Our results are probably overly optimistic, and at worst, they might be an illusion. Our results might be fooling us.

I'm skeptical of using backtested, in-sample results, used to discover monthly patterns, and expect good real-world results.

One way around this is to leave a segment of historical data as out-of-sample. For example, leave the last four years out of your in-sample backtest. Then locate your favored parameters and move the strategy to the three years of unseen data. The strategy results on the out-of-sample are what you would use to judge the quality of your strategy.

But I'm going to try something else.

I want to validate our results using a rolling window over the historical data. Thus, I will use TradeStation's Walk Forward Optimizer to test our idea further.

Testing on Walk Forward Optimizer

I first tested only taking trades in a bull market. Once again, I  optimize the "OpenDay" input with values 1-31 and the "HoldDays" input with values 0-4. Then used the optimizer to perform a cluster analysis.

The results are not optimistic. We can see many of the tests failed. 

I then performed the same test during a bear market and again, not looking so good.

I then decided to remove the regime filter and test once more. This looked better.

This had several more passing tests, but it was still far from ideal.

This is not to say we can't make a trading system from this concept. Some people may be trading something like this. Maybe that person is you. However, this makes me pause that such a simple system would work on the live market.

How would I proceed?

When I started working on this study, the results looked promising. However, the rolling window for optimizing the two input values threw a cold water on that hot idea.

This article is an excellent example of how we can fool ourselves into thinking we found an edge and this could lead us to trading live and losing money. Of course, maybe we do have a valid edge! But, when we tried a different method of analyzing the input parameters, the results don't look good. Thus, caution is warranted.

So, for moving forward, I would put this concept into incubation. Take the results of the Walk Forward Optimizer without the regime filter, which is the last WFO test we did. Select the optimal result, the five runs with 20% OOS. Here are the optimal parameters for the test:

OpenDay = 14
HoldDays = 0

These settings are the latest optimization that are valid through September 4, 2024. I would put this into incubation and see how it performs in 2023. That would be a good starting point.

The green color highlights trades that are averaging over $30 per trade. We can see the best opportunities occur at night on the long side! Shorting continues to be problematic in both the day and night sessions.

So what does this tell us?

These results demonstrate that:

1. It's much easier to build strategies on the long side during a bull market. Makes sense.
2. Opportunities may be most substantial in the overnight session.

Inverted Regime Filter 

Next, I want to perform this test again but invert the regime filter, so I will take long trades only in a bear market and only take short trades in a bull market. To perform this test, I used the EasyLanguage code provided in the previously mentioned article. So, review that article if you want a copy of the code. Below are the results.

Once again, the green color highlights trades averaging over $30 per trade. And once again, where is all the green seen? That's right, going long! You would think shorting in a bear market would be obvious. But it could be more apparent with this test.

So what does this tell us?

1. It's much easier to build strategies on the long side during a bear market. A bit surprising!
2. Opportunities may be most substantial in the day session. Notice this is the opposite of a bull market. 

Conclusions

For the U.S. stock index markets, the low-hanging fruit continues to be found on the long side. However, depending on the market regime, you will want to switch between the day or night session.

1. During a bull market, focus on overnight trading
2. During a bear market, focus on day trading

There is a difference between the day session and the night session. This difference is even more dramatic when adding a simple moving average regime filter and only trade trades during a bull market.

We want to go long during a bull market based on what we know about the E-mini S&P market. We also want to buy during a pullback.

Recent Performance Of The Overnight Edge

Let's take a closer look over the past couple of years. Below are the same tests performed from February 11, 2019 to February 9, 2022 with no slippage or commission cost deducted.

It really looks like the overnight edge is holding up well. When taking trades at night during a bull market, that equity curve looks like a trading system. But remember, this is a market study, not a complete trading system. These strategies have no stops and take a new trade every day.

What About Other Markets?

I took a look at the other major stock index markets, and they show similar behavior. Below are the results from taking long trades during a bull market in the overnight session. Click on the images for a larger view.

What Can I Do With This Info?

I think this information clearly points to a market edge. This would require holding a trade during the overnight session. What this study does not say is how to do that. Should you enter after the market closes? Maybe entering in the pre-market. Or, is holding all through the night the way to capture most of the points? Well, that will depend upon the market and the strategy you have.

Below are a few articles that point you in this direction to give you some ideas.

• S&P Overnight Trading Model
• Using Gap Zones To Improve The Overnight Edge
• Questions About The Overnight Edge
• Creating a Double Seven Intraday Strategy
• Building An Intraday RSI Trading System