Take profit condition on Daily P & L

[Kai]

I like to aim for a percentage profit per day and find I quite often meet it but then find it reverses to a lower profit or possibly a loss.

Sounds like it's just a vain attempt to keep some of that profit

If you are seeing patterns of going into profit and then going back to 0 or into loss there could be other reasons for this as Dallas mentioned, one of them could be that strategy has no edge

[ShaunWhite]

but I also believe in thinking outside the square, challenging norms and always looking for ways to improve processes.

Anyway, bed for me as I am into hospital for a non serious procedure on my back.
Good night.

All the best with the procedure, never a nice place to be.

Challenging preconceptions is great but bear in mind that gambling and probability have been studied for over 300yrs so there isn't much new anymore. And "trading" is just a straight bet on direction and size of move so all the usually value based decisions apply.

The biggest difference between what you're used to and sports is market independency, it's 100 unrelated markets a day not a few that rolls ever onward. So be mindful of the gamblers fallacy, if you have a strategy that fades during the day it's changing market types and conditions causing it because the markets don't know if you personally are up or down and treating you differently. Recalibrate for that change or eliminated the unfavourable markets and just keep grinding 247365.

[gazuty]

I use different automations on different markets as in: UK racing has a different automation to Aus racing and Aus racing country has a slightly different automation to Aus capital city racing.
From everything I have read on these forums the advise seems to be there is no such thing as a one automation fits all, so I don't quite understand what you are inferring.

Do you consider each racing event to be independent? Your answer seems to suggest that but what do you think?

[ShaunWhite]

From everything I have read on these forums the advise seems to be there is no such thing as a one automation fits all, so I don't quite understand what you are inferring.

What? I'm always banging on about one size fits all strategies. But although one of mine works on horses dogs UK and Aus, and potentially every selection, it doesn't always bet. But that's not dependent on what happened an abribary amount of time in the past be it 5ms 5mins or 5 weeks.

But market independence doesn't preclude different types of market. You can have strategy for each type if you want, large field small field etc, but the result of one race doesn't affect the result of the next, it's independent.

[gazuty]

But market independence doesn't preclude different types of market. You can have strategy for each type if you want, large field small field etc, but the result of one race doesn't affect the result of the next, it's independent.

Stop giving away the answer to the test in advance .

[ajanthony]

I use different automations on different markets as in: UK racing has a different automation to Aus racing and Aus racing country has a slightly different automation to Aus capital city racing.
From everything I have read on these forums the advise seems to be there is no such thing as a one automation fits all, so I don't quite understand what you are inferring.

Do you consider each racing event to be independent? Your answer seems to suggest that but what do you think?

Yes, I consider every event to be independent. The same as coin toss. Getting a head has no bearing on what the next flip will, it is still 50/50.

[gazuty]

Yes, I consider every event to be independent. The same as coin toss. Getting a head has no bearing on what the next flip will, it is still 50/50.

Do you consider you have an edge?

[ajanthony]

Yes, I consider every event to be independent. The same as coin toss. Getting a head has no bearing on what the next flip will, it is still 50/50.

Do you consider you have an edge?

I can see where you are going with the question. If I say yea I have an edge you will say I don't need what I am proposing. If I say no then logically you will say my system is flawed. Very logical.
I think I have an edge but haven't been at it long enough to give a definitive yes or no.
As I said previously, having a global P & L condition comes from my experience of other markets. The main advantage of it in those markets is that it levels out your equity curve and takes the volatility out of it. This can be proven in those markets because back testing on real data is possible, I am not aware this is possible on horse racing markets, maybe someone can show me otherwise.
I take in all the comments made in this post and it may well be the case it is not required but having the option (which Dallas has said is possible presently) is not going to hurt.
So, in the end, if it is currently possible we don't need to flog the subject anymore.
Thank you everyone for your input.
Anthony

[ShaunWhite]

This can be proven in those markets because back testing on real data is possible, I am not aware this is possible on horse racing markets, maybe someone can show me otherwise.

It's possible, many people store the api stream and run simulators and have done for over a decade. This is something lots of people have been doing as a fulltime job for 20yrs, if you can imagine it, people are doing it.

[ShaunWhite]

I don't think anyone has asked, what style of trading are you doing aj?

I understand why you'd want to stop if your strategy mid-session if it just isn't doing what you hoped. When people trial a strategy, which is what I think you're doing, the usual approach is to use very small stakes (except for speed critical in-running stuff as they take twice as long to place). Use a stake small enough that you're hedging for maybe +/-10p, and let it run for a few hundred markets, the longer the better.

Be prepared to spend a few quid gathering information about it, but as the markets are very efficient, unless your strategy is counter-intuitive then you should gross about breakeven. Then, once you have a database of bets (aka 'trades') that you consider to be a significant sample, several hundred maybe, start to analyse them for the situations that were slightly more advantageous. Refine and re-test, and repeat. When you're happy with it then very slowly start to increase your stakes over the course of about a month or so. You'll find a point at which your returns diminish as you start to affect the market.


You could of course instead go the whole 9 yards with data/simulators/back tests/servers etc but then you're looking at about a year before you're ready to go, unless you want to spend £200/mth per sport for data instead of accumulating it yourself. And you still need to find a strategy of course which isn't easy and could take another year, or might never happen.

The first suggestion is much cheaper and accessible for most newcomers, not many people buy a bike and want to wait a year to ride it. I'd always suggest starting simple and if you take to it, then you can always think about where you want to go.

[ajanthony]

I don't think anyone has asked, what style of trading are you doing aj?

I understand why you'd want to stop if your strategy mid-session if it just isn't doing what you hoped. When people trial a strategy, which is what I think you're doing, the usual approach is to use very small stakes (except for speed critical in-running stuff as they take twice as long to place). Use a stake small enough that you're hedging for maybe +/-10p, and let it run for a few hundred markets, the longer the better.

Be prepared to spend a few quid gathering information about it, but as the markets are very efficient, unless your strategy is counter-intuitive then you should gross about breakeven. Then, once you have a database of bets (aka 'trades') that you consider to be a significant sample, several hundred maybe, start to analyse them for the situations that were slightly more advantageous. Refine and re-test, and repeat. When you're happy with it then very slowly start to increase your stakes over the course of about a month or so. You'll find a point at which your returns diminish as you start to affect the market.


You could of course instead go the whole 9 yards with data/simulators/back tests/servers etc but then you're looking at about a year before you're ready to go, unless you want to spend £200/mth per sport for data instead of accumulating it yourself. And you still need to find a strategy of course which isn't easy and could take another year, or might never happen.

The first suggestion is much cheaper and accessible for most newcomers, not many people buy a bike and want to wait a year to ride it. I'd always suggest starting simple and if you take to it, then you can always think about where you want to go.

Hello Shaun,
I am running two types of automation. Scalping the first 3 favourites on UK races with conditions on number of selections, volume etc etc. I don't have a keep on unmatched bets and find that it is common for one of them to come in if the bet is unmatched and remains a straight bet.
The other is on Aus races backing the favourite but with conditions on volume and volume %, odds, number of selections etc.
I am collating all the data and so far I am getting good results on both automations. I run them live with small stakes but if I want to try a variation I test it on practise mode first so essentially I run live and practise in parallel but with variations on practise.
I have tried many automations and manual trading over the last year but this is the best results so far but I am not yet counting the chickens.
Whilst I have you can you help with this one. I want to lay the favourite but only once it is in running and its odds have gone out to around 10/1. I know I probably have to store a value of when it was favourite at 15 seconds pre off but I am not sure how to do it.
Can you offer a suggestion.
Thanks for your input above.
Regards,
Anthony

[ShaunWhite]

Whilst I have you can you help with this one. I want to lay the favourite but only once it is in running and its odds have gone out to around 10/1. I know I probably have to store a value of when it was favourite at 15 seconds pre off but I am not sure how to do it.
Can you offer a suggestion.

Sounds like you already know what you're doing with the testing, I'll stop being so patronising But it sounds like your strategies are straightforward enough to analyse in Excel with some data collected by BA and don't need any hardcore backtest setup.

I'm not the best person to ask about Guardian as I haven't used it for years but it sounds like you want to ...

Create a rule to Fix the order of favouritism at -15s
Create a Lay bet rule for selection 'by position = 1" that has the conditions "Inplay condition" = True and a Fixed Price condition of best lay price > 11 ?
(and the trigger time window will be up to you of course)

Something like that. I think it's just those two rules unless they've added bsp favouritism as a condition somewhere. There'll probably be more than one way to skin the cat.

[ajanthony]

Whilst I have you can you help with this one. I want to lay the favourite but only once it is in running and its odds have gone out to around 10/1. I know I probably have to store a value of when it was favourite at 15 seconds pre off but I am not sure how to do it.
Can you offer a suggestion.

Sounds like you already know what you're doing with the testing, I'll stop being so patronising But it sounds like your strategies are straightforward enough to analyse in Excel with some data collected by BA and don't need any hardcore backtest setup.

I'm not the best person to ask about Guardian as I haven't used it for years but it sounds like you want to ...

Create a rule to Fix the order of favouritism at -15s
Create a Lay bet rule for selection 'by position = 1" that has the conditions "Inplay condition" = True and a Fixed Price condition of best lay price > 11 ?
(and the trigger time window will be up to you of course)

Something like that. I think it's just those two rules unless they've added bsp favouritism as a condition somewhere. There'll probably be more than one way to skin the cat.

Thank you,
I will use your explanation on creating that rule and see if I can nut it out, if not, I will ask Guru Dallas for some help.
I must say my strategy is pretty simple, so excel should be fine. It makes me wonder though, is it too simple and will crash and burn, only time will tell.
Thanks.
Anthony

[gazuty]

So, in the end, if it is currently possible we don't need to flog the subject anymore.

One further suggestion. By all means implement your rule to shut down if profit reaches x or loss reaches y.

But, run in parallel in practice mode the same strategies on the same markets without the cut off and see what happens.

You will get further data and make better decisions.

It would be curious that the achievement of a particular amount of profit achieved from a particular point in time (the first race of the day) was a signal that no further profit is available on that day from the strategy. And if that were the case you would then have the data and be in a position to guess at how that very curious result came about.

[ajanthony]

So, in the end, if it is currently possible we don't need to flog the subject anymore.

One further suggestion. By all means implement your rule to shut down if profit reaches x or loss reaches y.

But, run in parallel in practice mode the same strategies on the same markets without the cut off and see what happens.

You will get further data and make better decisions.

It would be curious that the achievement of a particular amount of profit achieved from a particular point in time (the first race of the day) was a signal that no further profit is available on that day from the strategy. And if that were the case you would then have the data and be in a position to guess at how that very curious result came about.

Shall do.