Check my fat finger please

[sirastonb]

I wonder if someone can look at my Fat Finger file and see if its ok and if not adjust it or tell me whats wrong

[GunDog]

OK, normally @Dallas would jump in at this point and offer his advice. As a newbie I want to see if I actually understand automation, fortunately you have not said what this Fat Finger does? So I want to see if by looking the automation, can I tell what does?

Place a back bet on selection in position 1 @ 20 ticks above best market price for £5 one time with a rearm delay of 5 secs

Place a lay bet bet on selection in position 1 @ 20 ticks below best market price for £5 one time with a rearm delay of 5 secs

Then green all selections.

Am I correct so far?

[LeTiss]

For starters he's not actually rearming them, as they are instructed to place the bet just once

His back bet is all pre-race, but the greening up is once it's turned IP

His lay bet is supposed to last for 4mins after it's gone IP, but the greening up lasts for just 3mins

Sirastonb, there's an awful lot wrong with that automated strategy

[Euler]

The best way to create a rule is to put in one element, test it, then add another layer, test it, next one, test it.....

Eventually, you end up with an elegant rule.

[GunDog]

I take it that when you test a rule, it has to be done on an actual event?

[Emmson]

The best way to create a rule is to put in one element, test it, then add another layer, test it, next one, test it.....

Eventually, you end up with an elegant rule.

This will be my methodology, I am at year zero with automation but I fully intend to make strides through dedication and perseverance and arrive at an Elegant Rule

[ShaunWhite]

The best way to create a rule is to put in one element, test it, then add another layer, test it, next one, test it.....

Eventually, you end up with an elegant rule.

This will be my methodology, I am at year zero with automation but I fully intend to make strides through dedication and perseverance and arrive at an Elegant Rule

With all due respect that isn't a methodology that's just winging it. That works if you've got a ton of experience, but if it's new to you then if you just start knocking something together one line at a time without thinking about what's going to come later, you could get in a right mess, or need to go back and make loads of changes later. It might work eventually but it won't necessarily be elegant.

The first thing you want to do is switch off your PC and actually write down EXACTLY what you want it to do; when, how often, what conditions there are to place the bet, what conditions there are to close the bet, what conditions you need to stop it duplicating bets, when you might want to cancel unmatched bets, acceptable price ranges, staking method, re-arm times, max spread, what should happen when the rule runs a 2nd or 3rd time, bet persistance etc etc ec etc. EVERYTHING you can think of, not just in a quick minute or two but for a good 15 or 20 mins if you're new to it. There's no point even starting if you find out that once you think about it hard enough, it's illogical or you don't have access to what you need, or that it's way out of your depth for now. Try to end up with notes someone else could understand not just you, because otherwise you'll make assumptions you won't cater for.

At the same time, write a list of all the things you'll need to test for later such as what happens following a suspension, flip-flopping favourites, or races that go off especially late etc. Peter's suggestion works obviously, but it assumes you already have a pretty good idea about where it's going. Better to start with something a tad more formal, the point of a methodology is that you can repeat it, so perhaps if you had problems with a rule because you forgot to test something quite obscure (maybe withdrawn selections?), go back and add it to the list of things you want to test for and you'll never miss it again.

....and once it's written do spend a few mins documenting what your signal or stored value names are and what they do...you'll be amazed how you can look at something you did 6 months ago and forget what you were doing and why! So your methodology could be: devise, plan, write, test, document, backup, implement, refine. Then throw it all in the bin because it didn't work As a rule of thumb actual coding is usually only about 25% of a development job, less if the remainder of it has been done well and you don't need reworks.

I KNOW it all sound a bit joyless and OTT but LeTiss's comment a few post back is case in point, those aren't hard things to have thought about up front, but the OP has just put something together fast and not thought it all through, so is sat waiting for hours for some help. Once you're experienced you can just wing it but you'll make much quicker progress if you start with your stablisers on, because you'll spend much more time being productive if you're not trying to fix stupid errors you could have anticipated. Thing is people are usually in a hurry to do the exciting bit and make some money but like trading manually, you get better early results if you just slow down a bit and get a good grounding in the basics. You obvioulsy don't have to follow all of that to the letter but I hope there's one or two things in there that you might want to think about. All the best.

[Derek27]

The first thing you want to do is switch off your PC and actually write down EXACTLY what you want it to do;

Nobody uses pen & paper anymore, you need your PC.

[firlandsfarm]

....and once it's written do spend a few mins documenting what your signal or stored value names are and what they do...you'll be amazed how you can look at something you did 6 months ago and forget what you were doing and why!

I'm amazed if I look at something a week later and remember why I did it! And I agree Shaun, there's no substitute for writing it down, on paper … it seems to concentrate the mind more. I would go as far as recommending keeping a notebook/diary of your thoughts/activities … whenever you think of something new to try, note it down … when you think of something to change, note it down. If you change something hopefully for the better you must have a note of what was happening before, what you changed and what you expected to result from the change … without that you cannot make a comparison afterwards. And go back through your notes from time to time to stimulate more thought. It sounds kindergardenish but it will work if you persevere with it.

[napshnap]

....and once it's written do spend a few mins documenting what your signal or stored value names are and what they do...you'll be amazed how you can look at something you did 6 months ago and forget what you were doing and why!

I'm amazed if I look at something a week later and remember why I did it! And I agree Shaun, there's no substitute for writing it down, on paper … it seems to concentrate the mind more. I would go as far as recommending keeping a notebook/diary of your thoughts/activities … whenever you think of something new to try, note it down … when you think of something to change, note it down. If you change something hopefully for the better you must have a note of what was happening before, what you changed and what you expected to result from the change … without that you cannot make a comparison afterwards. And go back through your notes from time to time to stimulate more thought. It sounds kindergardenish but it will work if you persevere with it.

It's not kindergardenish at all. It's natural.

[MemphisFlash]

i checked my fat finger, works fine, lol

fat.jpg

[ShaunWhite]

The first thing you want to do is switch off your PC and actually write down EXACTLY what you want it to do;

Nobody uses pen & paper anymore;)

And that's their problem. With a notepad there's no distractions and you can also annotate more easily, and a notebook is so much easier to browse than a load of random files. There's going to be some personal preference or generational thing but operating a device while allegedly focusing (or listening to music at the same time ) means you're not 100% concentrating. I defy anyone to sit and actually get in the 'zone" of a problem for even 30mins in front of a computer without at some point doing or looking at something else. Given the type of problems you see here Derek, I'm not sure even operating a pencil at the same time as thinking is a good idea.

Firlandsfarm knows what I mean a whole different bunch of synapses start firing once you stop diddling around on your computer. We've only got 8-bit brains, it's why when people think really really hard they even close their eyes to try and reduce the inputs.

[Derek27]

The problem I have with pen & paper, firstly, it takes half-an-hour to find a pen and an unneeded piece of paper to write on, and secondly, my handwriting is now barely readable - it's so long since I put pen to paper besides ticking boxes.

[ShaunWhite]

The problem I have with pen & paper, firstly, it takes half-an-hour to find a pen

That's the problem
I was hunting for a pencil myself the other day and found this one.... google says the last time a telegram was sent was in 1981 ! For the youngsters, a 'telegram' was a text message delivered by a lad in a smart uniform I think it's from the first place I did some coding for, jeez time flys .

20200810_154413.jpg

[Emmson]

I take on board ShaunWhite's thoughts but I am not a newbie newbie just as regards automation which I've never done before, manual trading I'm very experienced with and no mug.