Trading What I see !?

[speedyhamster]

Believe that you will do something and you will do it

[goat68]

bot resuming it's losing ways... now -£6 down for the week...
Interesting, this plays into the previous few posts about psychology... i'm very much doubting my edge in my current bot method, lack confidence in it. So much so i'm going to tell you what it does...
Essentially it is monitoring market volume flow across the front favs in the last 4mins before the off. It triggers a trade when it gets a large enough volume spike, and it then enters on that selection in the direction of the volume momentum. My theory being it is more likely to continue than not, as the surge of volume is likely genuine money before the off. I have tried various exit and profit target exits, and stop loss methods. The current bot this week does not have a stop loss, and has a small take profit target order. So very similar to Peter's offset bet video in terms of trade management.
So i've watched quite a lot of the bot entries and noticed a number of behaviours:
- It triggers quite often on the 2min surge point
- It quite often triggers on those moves where you see the Betfair charts go wrong
- When it triggers the price typically stalls, sometimes it surges to my profit order, but quite often if reverses and wiggles for a while. Other times it reverses for a decent loss!
- Hit rate is reasonable 60%, but this week with no stop loss it's had a number of big losses compared to the profit target. I was hoping for 75%, which it hit on tuesday, but gone down hill since...

I'm not convinced there's any edge in trying to hop on volume money momentum pre-off.

I've had a number of msgs from fellow Automation members on here (thank you!), and there's an interesting theme, basically what BA Guardian has to offer is not really enough to develop a long term profitable automation method. Most have developed there own bespoke software.

[Trader Pat]

I've had a number of msgs from fellow Automation members on here (thank you!), and there's an interesting theme, basically what BA Guardian has to offer is not really enough to develop a long term profitable automation method. Most have developed there own bespoke software.

There are plenty of full time auto traders on the forum who use Guardian and make very healthy profits so don't fall back into blaming the tools and not the workman.

Ultimately though Goat you could have saved yourself a lot of time and effort from the start if you would have said that your strategy was basically to jump on the back of a volume spike. Unfortunately your theory is wrong, a spike in volume can set off a massive move, or it can cause a move of a few ticks or sometimes it can set off a move in the opposite direction, it all depends on many factors.

The trick about trading momentum is you have to identify if its real momentum and if it is then you have to identify when its coming to an end.

[alexmr2]

Might be interesting to post this here, my 3 months P&L vs this time last year. I've came to the conclusion if you keep practising you can't get worse obviously some very stupid mistakes inbetween but hopefully this year it will finally reverse

[goat68]

I've had a number of msgs from fellow Automation members on here (thank you!), and there's an interesting theme, basically what BA Guardian has to offer is not really enough to develop a long term profitable automation method. Most have developed there own bespoke software.

There are plenty of full time auto traders on the forum who use Guardian and make very healthy profits so don't fall back into blaming the tools and not the workman.

Ultimately though Goat you could have saved yourself a lot of time and effort from the start if you would have said that your strategy was basically to jump on the back of a volume spike. Unfortunately your theory is wrong, a spike in volume can set off a massive move, or it can cause a move of a few ticks or sometimes it can set off a move in the opposite direction, it all depends on many factors.

The trick about trading momentum is you have to identify if its real momentum and if it is then you have to identify when its coming to an end.

I was just pointing out a correlation of those that have PMsged me...

I thought my theory was wrong based on the results so far, but i'd just like to argue a point you've hightlighted, in that my strategy is flawed, yes, yet people will keep on saying you need a larger sample size... which maybe true for a profitable strategy, but please be wary of saying just keep going and lose more money...! Luckily -£50 so far this year is nothing.

[goat68]

Might be interesting to post this here, my 3 months P&L vs this time last year. I've came to the conclusion if you keep practising you can't get worse obviously some very stupid mistakes inbetween but hopefully this year it will finally reverse

Blimey, that's some trend, hope you turn the corner soon

[goat68]

Another observation i've noticed watching some of my bot entries, is it's quite often Layed near cross-over points, good yes? well not when it has done it.... straight through it goes...!
Which does beg me to think, so we're at a crossover point, everyone and his dog knows that, so they all start laying, hence volume surge to lay, this just plays into the pro-traders who offer lower and lower prices.....

[goat68]

I'm going to let the current bot run the week out... then bin it.
I'm going to change to work on a new bot purely looking at the historic fav prices, as per Peter's videos... Horse A is coming in, horse B and C are heading out, B & C are at key prices, so we'll try backing A...ie.being more "anticipating", rather than trigger off some technical event.

[goat68]

I do feel i'm probably flogging a dead horse. I really don't have the years of manual experience as a background to developing an automation strategy that might work, my day job means I cannot manually trade any more, so tricky one. I'm a decent programmer, and like coding, so will continue bots for a while, but I think it might lose it's attraction if it's not producing any results....Alexmr2's 2 year trend is a great example.

[Trader Pat]

I was just pointing out a correlation of those that have PMsged me...

I thought my theory was wrong based on the results so far, but i'd just like to argue a point you've hightlighted, in that my strategy is flawed, yes, yet people will keep on saying you need a larger sample size... which maybe true for a profitable strategy, but please be wary of saying just keep going and lose more money...! Luckily -£50 so far this year is nothing.

I'd never advocate needing a larger sample size for a strategy like this.

A volume spike is an important indicator that every trader uses but whether or not it sets off a big move is dependent on a host of other factors.

[speedyhamster]

bot resuming it's losing ways... now -£6 down for the week...
Interesting, this plays into the previous few posts about psychology... i'm very much doubting my edge in my current bot method, lack confidence in it. So much so i'm going to tell you what it does...
Essentially it is monitoring market volume flow across the front favs in the last 4mins before the off. It triggers a trade when it gets a large enough volume spike, and it then enters on that selection in the direction of the volume momentum. My theory being it is more likely to continue than not, as the surge of volume is likely genuine money before the off. I have tried various exit and profit target exits, and stop loss methods. The current bot this week does not have a stop loss, and has a small take profit target order. So very similar to Peter's offset bet video in terms of trade management.
So i've watched quite a lot of the bot entries and noticed a number of behaviours:
- It triggers quite often on the 2min surge point
- It quite often triggers on those moves where you see the Betfair charts go wrong
- When it triggers the price typically stalls, sometimes it surges to my profit order, but quite often if reverses and wiggles for a while. Other times it reverses for a decent loss!
- Hit rate is reasonable 60%, but this week with no stop loss it's had a number of big losses compared to the profit target. I was hoping for 75%, which it hit on tuesday, but gone down hill since...

I'm not convinced there's any edge in trying to hop on volume money momentum pre-off.

PnLChart2.jpg

This makes sense to me the graph, from my experience of volume spilkes you need to trade what you see, therefore a servent would be needed if wanted to automate this

[speedyhamster]

I was just pointing out a correlation of those that have PMsged me...

I thought my theory was wrong based on the results so far, but i'd just like to argue a point you've hightlighted, in that my strategy is flawed, yes, yet people will keep on saying you need a larger sample size... which maybe true for a profitable strategy, but please be wary of saying just keep going and lose more money...! Luckily -£50 so far this year is nothing.

I'd never advocate needing a larger sample size for a strategy like this.

A volume spike is an important indicator that every trader uses but whether or not it sets off a big move is dependent on a host of other factors.

Agree

[ShaunWhite]

You can certainly make money using Guardian, what I meant was that it's harder if you only use Guardian, and for certain techniques it's not as adept as a piece of bespoke software would be. That's understandable because Guardian is an incredible Jack of all trades, not just tuned for one specific job.

Sample size etc : As I've mentioned you'll never get a big enough sample just running stuff in live, you'll go broke before you do. Use your skills to harvest data, and begin to refine your strategies using that. In six months you'll have a data set that's definately a large enough sample for anything, and you can try ideas and prove and disprove things to your heart's content. You can even start to get some sort of data together using BA/Excel without having to dive straight into the API, although that's probably where you'd want to go eventually for scale and granularity.

...Take your current strategy,

• What's a spike anyway? Is it relative to the expected market total? Is it relative to the prev 2 mins vol? Is it relative to the price/vol? How about maidens vs handicaps?
  
  Does a spike cause shortening, lengthening, with the spike direction, against it.....and is that just for the next 10s or the next 30s, or to BSP?
  
  Do stop losses and take profits have any advantage/disadvantage, and if so then what %age or maybe it should be a time stop not cash?
  
  How about staking? Level or By Liability?
  
  All runners? Or just the fav? Or everything priced below X?
  
  Backs, lays, or either, and do you flipflop positions on a second spike the other way?

You will NEVER answer those questions without decent data, for no other reason than the number of permutations and parameters would take thousands of years to cycle through.....and then picture trying to modify/tune that strategy for dogs, horses, uk, aus, win, place, or maybe football. Do you really want to do that in live? Or do you just want to drop another dataset in the same model and press a button, or leave 20 variations running overnight and wake up to the results.

I'm NOT saying that people can't make money using Guardian without doing all that, they do, but those that do usually have a pretty solid idea of what they know works and are looking to scale up if they can't physically exploit every situation, or just want to execute more efficiently. You need to bring everything you've got to the trading table and work to those strengths, you're clearly not a devotee of the gambling world but you know a bit of tech, so use it rather than trying to become an expert on punter psychology and their tendency (or not) to follow trends. It's obviously going to need some effort and hard work, so it's down to you and how you view that investment/reward equation and your estimate of your odds of success with/without.

Long one, sorry about that, but the pros and cons and details would need many long nights in the pub to get through. Anyway I'm not here to make you do anything you don't want to do, but I do think you need to open your eyes to how hard edges are to find, especially the way you're choosing to find them. It won't guarantee success but imo it'll take you from a 1% chance to maybe a 50% chance.

[ShaunWhite]

btw your chart: red was probably a dog (maybe not, it could just have greater natural varience) but for the others, I don't think whatever you've been changing has had any effect. It's probably an irrelevent parameter(s) as they're all very similar. The only way to tell if a parameter is relevant is to cycle through a range of values and see if the PL from each iteration forms a bell curve, ie a sweet spot and tails as you move away from it. The chart actually looks like a gross zero edge that's declining by commission, which is exactly what you'd expect with your 'random' strategy.

[beermonsterman]

Some good reading their I missed out on yesterday good to see people jumping in to give some pointers when someone is trying hard but feels like they not getting anywhere keep going Goat you will crack it eventually I may be an exception to others due to how long I've been doing this and still not profitable but I know this I may not be a fast learner but my perseverance is now starting to pay off things are coming together at last and I believe they will for you too