Bot Trigger Building Part 2

Now that the selection has been identified (in Part 1, a horse in lane x in a race with more than x selections), I'd like some triggers to be met before initiating a trade.

Firstly, for this example, I'd like to place a lay bet only when the horse has odds in range 10 to 20 when the following 2 triggers are met:

1- Our selection has dropped in position since X seconds in play.

2- The number 1 and 2 position horses both have back odds less then 2 (and our selection is not position 1,2 or 3)

So after Part 2, we have a working bot that will execute a lay bet within odds range of 10 to 20 on our selection (based on the 2 triggers in Part 1) ONLY IF the 2 triggers in Part 2 are true

Comments ( 19 )

  • Stefan
    27.9.2019 18:44:20

    How is evaluated “…our selection has dropped in position...”?

    27.9.2019 23:57:02

    After for example 20 seconds in play, I would like to check the position of the horse (using it's odds). When (if) 2 horses achieve odds less then x ( 2 for example), then re-check the position of our selection and compare it to the position it was at at 20 seconds. So if at 20 seconds, it is at position 2 and then when trigger is 2 horses are at odds below 2, it re-checks the selection and sees that it is at position 5, the selection has dropped in position.

  • Stefan
    28.9.2019 0:46:54

    Ok, but you did not specify that you want to check horse's positions in two different times. My first implementation was that bot waits till market is turned at in-play, then saves information about all allowed selections, so horse's positions is saved as well.

    At set time in race, bot checks your conditions, so favourites must be traded at that moment of checking under set price/odds.

    Then allowed selections are sorted by position change difference, so biggest position change is the first one, if this selection is not among the first three favourites, it is nominated as the selection your strategy will be executed on.

    The bot trigger again similar like I explained to you in the part 3, does not have to know anything about your criteria for entry point, so if you allow starting lay/back trading session only when the selection/horse is traded from 10.0 to 20.0, then you can simply set that to Entry Criteria for this lay trading bot.

    Now think a little bit. I hope you already know that Entry Criteria could be evaluated just once, or could be evaluated till all set Entry Criteria are met. This feature gives you two different options for your bot placing a lay bet.

    If you think more, then you will find that bfexplorer offers more bots placing a bet, or starting trading session, namely:

    Place Bet

    Be the First in Queue

    Fill or Kill

    Place Bet and Close Selection Bet Position

    Scratch Trading

    This gives you really big flexibility for different ways your bet is placed on the market selection, and closes your bet position for loss when needed.

    Of course not all of above bot strategies to place a bet are suitable for in-play markets where odds changes so rapidly as on the horse racing, therefore I would suggest to use:

    Place Bet and Close Selection Bet Position

    But of course you can test them all, or other combinations for instance:

    Place Bet, and Trailing Stop Loss

    What is very good news for you is that bot trigger code does not have to implement anything to place a bet and manage its status, so you will get simpler trigger code.

    I really cannot imagine how people could manage such strategies just in Excel spreadsheets, so traders using Bet Angel, Cymatic or Gruss software. I think they cannot run such strategies, am I right?

    Your tutorial in the part 1 is ready, I will post bot trigger code for the part 2, but first test your bot trigger for part 1, and as the simple coding exercise, change the bot trigger code to report not only horse name, but the last traded price as well.

    Here is the hint:

    30.9.2019 2:52:29
    The following is the Part 1 code with the last price traded added. In Part 2, we will build triggers into this that react to our selection position change from starting point (x seconds in play) to trigger point (The number 1 and 2 position horses both have back odds less then X (and our selection is not position 1,2 or 3) module BfexplorerBot //(* #I @"C:\Program Files (x86)\BeloSoft\Bfexplorer\" #r "BeloSoft.Data.dll" #r "BeloSoft.Betfair.API.dll" #r "BeloSoft.Bfexplorer.Domain.dll" #r "BeloSoft.Bfexplorer.Trading.dll" #r "BeloSoft.Bfexplorer.Service.Core.dll" //*) open System open BeloSoft.Data open BeloSoft.Bfexplorer.Domain open BeloSoft.Bfexplorer.Trading let toIndexes (value : string) = try value.Split(',') |> Int32.Parse |> List.ofArray with | _ -> List.empty /// /// HorseRacingBotTrigger /// type HorseRacingBotTrigger(market : Market, _selection : Selection, _botName : string, botTriggerParameters : BotTriggerParameters, _myBfexplorer : IMyBfexplorer) = let allowedSelectionIndexes = toIndexes (defaultArg (botTriggerParameters.GetParameter("AllowedSelectionIndexes")) "1,2,3") let isHorseRacingMarket() = market.MarketInfo.BetEventType.Id = 7 && market.MarketDescription.MarketType = "WIN" let getActiveSelections() = market.Selections |> Seq.filter isActiveSelection |> Seq.toList let getMySelectionsData() = maybe { if allowedSelectionIndexes.Length > 0 then let selections = getActiveSelections() let mumberOfSelections = selections.Length let maximalIndex = allowedSelectionIndexes |> List.max if mumberOfSelections >= maximalIndex then return allowedSelectionIndexes |> (fun index -> selections.[index - 1]), mumberOfSelections } interface IBotTrigger with member __.Execute() = if isHorseRacingMarket() then match getMySelectionsData() with | Some (selections, mumberOfSelections) -> let mySelectionNames = selections |> (fun mySelection -> mySelection.Name) |> String.concat ", " let mySelectionBookValue = selections |> (fun mySelection -> mySelection.LastPriceTraded) |> List.sum TriggerResult.EndExecutionWithMessage (sprintf "\nMy selections: %s %f\nNumber of runners: %d" mySelectionNames mySelectionBookValue mumberOfSelections) | None -> TriggerResult.EndExecutionWithMessage "Failed to make my selections!" else TriggerResult.EndExecutionWithMessage "You can run this bot on a horse racing market only!" member __.EndExecution() = ()
  • Stefan
    30.9.2019 9:56:56

    Peter, I do not want to patronize you, but I have got very simple question?

    Are you satisfied with what you did?

    Why did I ask?

    Because any written source code must be formatted to be readable by other programmers.

    Yes, some programming languages that use for instance { } to separate block of code, could be stripped to one line and compiler successfully build such code, but not F# where code block must be formatted by text/code indentation and you cannot used tabs:

    I mentioned that, because during weekend you posted video showing you edited bot trigger in Notepad text editor. So anyone who would like to use your code you posted in your comment, must copy and paste your text, reformat it correctly, and then will find out that your code does not do what you claim you did.

    So you waste maybe 10 minutes or more, to someone who wants to test your code. English is foreign language for me so is very hard for me to find politically correct words for your behavior.

    Please, install either Visual Code, like I show here:

    Betfair Bot Programming for Non-Developers

    Or Visual Studio Community version:

    Bfexplorer - Betfair BOT SDK

    Both IDE (Integrated Developer Environment) applications are free, unlike Microsoft Excel other betfair app use, where you must buy licenses of Microsoft Excel. Using Bfexplorer BOT SDK all applications or tools helping you to create bot trigger are for free.

    Why is using right programs/applications important when developing/writing code?

    Because like text editors with spellchecking, such IDE checks your code when writing, so any mistake in misspelling anything in code is reported. Visual Studio, or Visual Studio code offer you as well, so called IntelliSense/IntelliCode technology when writing code, so helping you to write the code:

    And finally why did I ask you to make some changes in the code?

    To force you learn something, because bot trigger code in other parts of our tutorial will be more difficult.

    I know you are not software developer, but in part 1 the source code is really readable even by non programmer, it contains plain English words like if, then, else, match, let, try, with, and some words you would maybe need to think about, like open, type, interface, member and map

    Please, when you want to use source code, and want to share it with others, do them favor and share it on through Github Gist, like I do:

    Again, you can register on for free. Or just use any of other code sharing services on the Internet:

    30.9.2019 13:51:49

    Sorry about the formatting.  I copied and pasted the working code and it looked good on my side with the correct indentations etc,, but the website changed it to the posted result that is there now.  I'll look at setting up GitHub and reposting the link to the code so others who are following/learning can view and get it.

  • Stefan
    30.9.2019 16:21:59

    Here is my first test that actually triggered some trading session. When testing it is really important how you set all trigger parameters, for testing it is good to set them without limits, so not parameters are tested but code

    I set the fallowing parameters as default ones:

    allowedSelectionIndexes: 1,2,3,4,5,6

    timeToStart: 40.0

    favouritesPrice: 50.0

    When the race started, bot trigger saved position of allowed horses. When timeToStart expired, all allowed horses were evaluated again

    positionDifference <- startPosition – position

    So trigger selected Bianca Minola as nominated selection on which Bot Trigger started lay/back trading session for 4 ticks of profit.

    1.10.2019 8:59:15

    I am not a programmer, but this is the part 1 with the Last Price Traded added to output

    I am looking forward to viewing Stefan's code with Part 2 combined into Part 1 :)

  • Stefan
    1.10.2019 9:21:15

    Ok, it seems you managed to understand the concept of F# programming.

    On the other hand I see you have actually no understanding of betting terminology, as book value is not calculated as you did in your code. The book value in betting is sum of selection’s probabilities.

    Here is my bot trigger code for part 2: HorseRacingBotTrigger2.fs

    1.10.2019 10:01:51

    'F' grade for me this time.  Thanks for the solution to Part 2 :)    What homework would you like me to submit before the final reveal in Part 4? - I hope to redeem myself