BenchMark_ET.xls measures Excel performance with 6 trading related Excel/VBA tests. Once the benchmark is complete, your results can be submitted to the Database (just press “Submit”). All submitted results are available here (all results are sortable). The default view is most recently submitted result topmost (so your result should be on top until you sort). System Information is listed along with each test result (processor name, OS, Clock Speed, Cores, Excel Version, File type). A composite Scoring system is in place, but each individual test result is also listed in the results database. All results can be sorted in ascending/descending order by clicking each result’s title. For best results disable autosave and background error checking in Excel options.

Instructions:

1. Download Benchmark_ET.xls.
2. Open Benchmark_ET.xls and press the button “Start Benchmark”
3. (optional) Submit and view your results.

Below is an explantion of each test:

Test 1: Data Build Time

Measures the time it takes VBA to create 5 columns * 65535 rows of random tick data.

Test 2: MACD VBA Calculation

Measures the time it takes VBA to calculate the MACD indicator for the 5 columns * 65535 rows of data created in test one. (The VBA code for the MACD calculation is here.)

Test 3: Dispayed Ticks (30 Sec)

Measures the number of ticks (price changes) than can be displayed in 30 seconds. In my original benchmark this was the only type of test. The results varied so greatly that newer versions of excel took hours to complete the benchmark while older versions took just minutes making the benchmark annoying to run. By limiting this test to 30 seconds the total time to run the benchmark will be similar among all systems tested.

Test 4: 63000 ticks converted to OHLC

After 63000 ticks are created, this test converts the tick data to OHLC data. The result is measured in time.

Test 5: Live formula Test

This test begins with with 5000 bars of OHLC and a series of formula based indicators. The last price changes in E5000 as rapidly as possible but waits for all formulas to calculate between each price change.  The benchmark result is the number of price changes that can be completed with formula recalculations within 30 seconds.  Below is the list of indicators calculated with live formulas in this test.

1. SMA (100,200,300)
2. RSS
3. RSI
4. Fisher Transform of RSI

Test 6: Live formula Test

This test is identical to test 5 except that all formulas also refer to and are dependent on the changing cell, E5000. Test 6  compared to Test 5  is meant to test the “smart recalculation engine” promoted by Microsoft.  More info can be found in this Microsoft article under the section titles “Understanding Calculation Methods in Excel”.

Scoring:

A composite scoring system is in place and the result is listed in column two on the results data base page and also at the top of the benchmark results form upon completion. With the exception of test 3 all tests are counted equally. Test 3, which tests displayed ticks on a stock chart is counted 3 times because for trading it’s an important result to consider. To establish the composite score 20 tests were run on various systems and excel versions. The best and worst scores for each test were used to establish the benchmark score. The best scores from the initial tests resulted in a score of 100 and the worst scores resulted in a score of 0 therefore scores greater than 100 and below 0 are possible.  On the results page idle processor speed is shown and overclocked systems may not show the true processor speed. If you see two systems that look the same with significantly different scores, the higher score is likely from an overclocked system. Thanks for trying my Excel benchmark and submitting your system’s result!

As an example, SPY futures are available in contracts with a multiplier of 1000 and 100 on the ONE exchange.

The error message received in this case is “The contract description specified for SPY is ambiguous;”

The solution is to first find the local symbol. You can do this by first adding the contract in TWS, right click on the contract name within TWS, then select contract details and select “details”.
This will open the page shown below and you can find the local symbol. As shown below, leave out the other parameters such as expiration date as it’s inherit in the local symbol (and will result in no data).

Both files contain the exact same set of Open High Low Close data. The formula based version is easier to understand and serves as a way to verify that the VBA code that I wrote is correct. Both methods give the same result.  The major benefit of using VBA is that the parameters for MACD can be easy changed from the input boxes. In addition the VBA method shows only the final result rather than take up five columns.

The VBA based version can be downloaded here.

The formula version can be downloaded here. Below is the code from the VBA based version.

Sub ETmacd() 'written by Exceltrader www.exceltrader.net
Dim EMAslow As Double, EMAfAst As Double, ws As Worksheet, LR As Integer
Dim eMaF() As Double, eMaS() As Double, EMAdif(), emaPer() As Double, MacDper As Double, coUnt As Integer
Dim DataRange As Range
Dim ExPSlowWeight As Double
Dim ExPFastWeight As Double
Dim PerWeight As Double
'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
''  The below three lines are the MACD settings.                                         ''
''  The Values can either be changed here or uncomment the inputbox lines to be prompted.''
'''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''''
EMAslow = 13 'InputBox(Prompt:="Enter Macd Slow settings.", Title:="MACD SLOW", Default:="13")
EMAfAst = 5 'InputBox(Prompt:="Enter Macd Fast settings.", Title:="MACD Fast", Default:="5")
MacDper = 6 'InputBox(Prompt:="Enter Macd Period settings.", Title:="MACD Period", Default:="6")
ExPSlowWeight = 2 / (EMAslow + 1)
PerWeight = 2 / (MacDper + 1)
ExPFastWeight = 2 / (EMAfAst + 1)
Set ws = ThisWorkbook.Worksheets("VBA") 'or use exact sheet name for example ThisWorkbook.worksheet("Sheet1")
'slow
    With ws
    LR = .Cells(Rows.coUnt, "A").End(xlUp).Row
        For Each DataRange In ws.Range(.Cells(2, "A"), .Cells(LR, "A"))
        coUnt = DataRange.Row + 1
        'fill the eMA slow Array
        ReDim Preserve eMaS(1 To coUnt)
                If coUnt = EMAslow + 1 Then
                    'get the first value which is the Simple Moving average
                     eMaS(coUnt) = Application.Average(ws.Range(.Cells(2, "E"), .Cells(coUnt, "E")))
                ElseIf coUnt > EMAslow Then
                    eMaS(coUnt) = (.Cells(coUnt, "E") * ExPSlowWeight) + (eMaS(coUnt - 1) * (1 - ExPSlowWeight))
                End If
        Next DataRange
'fast
        For Each DataRange In ws.Range(.Cells(2, "A"), .Cells(LR, "A"))
        coUnt = DataRange.Row + 1
        'fill the eMA slow Array
        ReDim Preserve eMaF(1 To coUnt)
                If coUnt = EMAfAst + 1 Then
                    'get the first value which is the Simple Moving average
                     eMaF(coUnt) = Application.Average(ws.Range(.Cells(2, "E"), .Cells(coUnt, "E")))
                ElseIf coUnt > EMAfAst Then
                    eMaF(coUnt) = (.Cells(coUnt, "E") * ExPFastWeight) + (eMaF(coUnt - 1) * (1 - ExPFastWeight))
                End If
        Next DataRange
        ReDim Preserve EMAdif(EMAslow To UBound(eMaF))
            For coUnt = EMAslow + 1 To UBound(eMaF)
                EMAdif(coUnt) = eMaF(coUnt) - eMaS(coUnt)
            Next coUnt
'MacD Period
        Dim x As Integer, y As Integer, z As Integer, Avee As Double
        y = EMAslow + MacDper - 1
            For x = y To UBound(EMAdif) - 2
                'get the SMA for first value
                If x = y Then
                        For z = EMAslow + 1 To EMAslow + MacDper  '(EMAslow + MacDper - 1)
                        Ave = Ave + EMAdif(z)
                        Next z
                        ReDim emaPer(x To LR)
                        emaPer(x) = Ave / MacDper
                ElseIf x > y Then
                emaPer(x) = (EMAdif(x + 1) * PerWeight) + (emaPer(x - 1) * (1 - PerWeight))
                End If
            Next x
  x = Empty: y = Empty: z = Empty
  x = LBound(emaPer)
  y = UBound(emaPer)
  For z = x To y - 1
  .Cells(z + 1, "J") = EMAdif(z + 1) - emaPer(z)
  Next z
End With
End Sub 'written by Exceltrader www.exceltrader.net

DataConverter.xls allows you to configure the Number of OHLC bars that should be condensed down into a single bar. You also can list the file name, sheet, and columns where your data is located.  The data will almost instantly be transformed into the higher time frame you have requested.

Instructions:

1. Download DataConverter.xls (*last updated 1/1/2011)
2. Open DataConverter.xls. (Quickstart!: Skip the directions below. Just Open file and press “Convert” to see how it works with default settings.
3. If you have OHLC Data in a workbook, open that file (you can also just use the default sample data that is contained on the sample data sheet in DataConverter.xls)
4. Enter the Column letter for each type of data (Time/date, Open, High, Low, Close, Volume)
5. Enter the Name of the WorkBook and worksheet that contain the data. For example, to convert 1 Minute bars to 4 minute bars, simply enter 4. To convert 3 minute bars to 15 minute bars enter 5 (5*3=15) etc etc.
6. Finally press the “Convert” Button.  The converted data will appear in a new workbook. Save as desired.

The constant volume bar option, updates a new bar after a certain amount of volume has completed. This amount should be entered into D3 of the “Live” sheet. Checks are in place to prompt if the values are not entered and the constant volume is set to override the default time based bars.

Below is an example of Constant volume bar charting. You will notice that each bar contains roughly the same amount of volume.

A little History

Over the last 16 years of trading, I have used a lot of trading platforms: broker platforms, third party vendor platforms and small applications written by private programmers. One thing all these platforms have had in common is that at some point I had a simple task or feature that was not already built in or impossible to implement.

As a long time user of Microsoft Excel, Excel became an obvious choice once Interactive Brokers began supporting excel with their API. I am not a programmer by formal education; however the learning curve for programming VBA for Excel is very fast. The key for me was buying a few great Excel books and using them for reference.

Why Excel may not be the best choice:

• Even standard (free) broker platforms a capable of a lot. If your trading strategies are not complex for example you only want to be able to enter stop losses and limit orders before going to work, there would be little reason to use excel as a trading platform.

• Excel can be a little too slow for scalping (entering and exiting within seconds), occasionally I have seen lag of a few seconds before excel begins sending the order.

• Although the learning curve is quick it does require a lot of time to program, error proof and test your work.

• For non Windows users or users wanting to change to another operating system, Excel would not be a likely choice.

Why Excel may be the best choice:

• Most people have at least some familiarity with Microsoft Excel. Any previous experience greatly decreases the learning curve.

• You can do almost anything in Microsoft excel. There has never been anything that I wanted to do in Excel that was impossible. I may have initiallythought it was impossible, but in the end I always found a way.

• There are fantastic resources to get help writing code. I have often asked questions on the Mr. Excel forums and received fast answers from professional programmers.

• Contrary to what many believe, you can create great charts in Microsoft excel. (The charts I post on this blog are in Excel!)
  Save Money! If you already own it then you save a lot of money by not having to pay for expensive trading platforms or charting programs

Is it the best choice?
For someone with little or no programming experience but who wants to implement complex trading strategies with a high degree (limitless?) of complexity and customization Excel may be the perfect choice. It certainly has been for me.