I dont think that'd be a problem, tho I dont know for sure. I do it 500 times an hour without a problem.goat68 wrote: ↑Sat Sep 25, 2021 10:01 amSo exactly, i would be doing it on every single bet, so I'd call that habitually!ShaunWhite wrote: ↑Sat Sep 25, 2021 5:11 amThey explain how to do it in the API FAQ so it's not a big deal, so long as it's used reasonably rather than habitually.
Trading What I see !?
Do you use BA though?Anbell wrote: ↑Sat Sep 25, 2021 10:24 amI dont think that'd be a problem, tho I dont know for sure. I do it 500 times an hour without a problem.goat68 wrote: ↑Sat Sep 25, 2021 10:01 amSo exactly, i would be doing it on every single bet, so I'd call that habitually!ShaunWhite wrote: ↑Sat Sep 25, 2021 5:11 am
They explain how to do it in the API FAQ so it's not a big deal, so long as it's used reasonably rather than habitually.
Im just hypothesising they would be more lenient to the BA subscription... since they know it's a standard feature of BA and BA is no doubt a significant volume them..
I'm not a normal user, but I suspect that you're seeing ghosts where there arent any. (I dont know though)
I follow Peter's twitter feed, just noticed this link today: https://www.betfairtradingblog.com/valu ... r-trading/
Of course I am doing "value betting", and this is a great article summarising my joys and woes, this paragraph is great and VERY true:
"
When you are doing well with a result dependent strategy, it’s easy to think you are a god and when you are doing badly it’s easy to think you are an idiot. This is one key reason that people have trouble replicating or even creating successful strategies.
"
Just wondering if Peter's reason for re-posting this was my recent Ups and Downs...!?!
Of course I am doing "value betting", and this is a great article summarising my joys and woes, this paragraph is great and VERY true:
"
When you are doing well with a result dependent strategy, it’s easy to think you are a god and when you are doing badly it’s easy to think you are an idiot. This is one key reason that people have trouble replicating or even creating successful strategies.
"
Just wondering if Peter's reason for re-posting this was my recent Ups and Downs...!?!
-
CallumPerry
- Posts: 575
- Joined: Wed Apr 19, 2017 5:12 pm
- Location: Wolverhampton
I’ve been thinking about some of the analysis comments from earlier in this tread for about 10 days now to the point where I feel inclined to ask more because I feel it may be really useful for my personal analysis.
I’m intrigued by the taking of the cumulative PnL and comparing it to an indicator such as the price-vwap. I believe it was said this checks the effectiveness of whatever you use as the x-axis i.e. price-vwap? I caught something about putting results into buckets and then doing this cumulative thing, I have no idea?
I’m mainly referring to Shaun’s explanation here at: Sat Sep 18, 2021 9:50 am
I’m theorising capturing a bunch of variables with each trade and creating one of these graphs for each to check their effectiveness simultaneously, not sure if this is how it works. Shaun or others, could anybody help me out please?
For (a bad) example, say I want to check whether my bot performs better at the point the markets has a high overround or a tight one, could I capture the book % at the point each of my trades close along with that trade’s PnL and group this data into discrete chunks i.e. book% = 100%-101%, 101.1%-102%, 102.1%-103% … 110%+
Then, map out the cumulative frequency or something and that will show where book% plays the most significant part in my system? If the results are similar with both halves of my sample, then it may be significant? Is that the rough gist of what’s going on? Apologies if I have misunderstood.
I’m intrigued by the taking of the cumulative PnL and comparing it to an indicator such as the price-vwap. I believe it was said this checks the effectiveness of whatever you use as the x-axis i.e. price-vwap? I caught something about putting results into buckets and then doing this cumulative thing, I have no idea?
I’m mainly referring to Shaun’s explanation here at: Sat Sep 18, 2021 9:50 am
I’m theorising capturing a bunch of variables with each trade and creating one of these graphs for each to check their effectiveness simultaneously, not sure if this is how it works. Shaun or others, could anybody help me out please?
For (a bad) example, say I want to check whether my bot performs better at the point the markets has a high overround or a tight one, could I capture the book % at the point each of my trades close along with that trade’s PnL and group this data into discrete chunks i.e. book% = 100%-101%, 101.1%-102%, 102.1%-103% … 110%+
Then, map out the cumulative frequency or something and that will show where book% plays the most significant part in my system? If the results are similar with both halves of my sample, then it may be significant? Is that the rough gist of what’s going on? Apologies if I have misunderstood.
-
ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
That's pretty much it Callum. I won't do a long reply with examples etc cos my pc is dead, and there's obviously nuance, but sorting results by things other than date can show where you're losing/winning. Grouping by bands is useful if you don't have a very big sample but you can usually get a feel for it not grouping. Your bet history and all the circumstances at the time the bets are placed is a really useful source of info.
-
CallumPerry
- Posts: 575
- Joined: Wed Apr 19, 2017 5:12 pm
- Location: Wolverhampton
Is this the right kind of way to set it up? I'm selfishly taking over the thread for my own development but I feel this could help Goat68 too. I hypothetically created a book% figure which would be captured as the trade opens in column A and a hypothetical PnL result -£2 to £3, over 1,000 trades, in column B. Columns G:L are the buckets and columns P:U contains the cumulative frequency.
As all of the results have a +ve expectancy, the graph shows all lines rising nicely. Is this how you would use this? Any series that goes -ve is a weakness of that strategy? Just before sending this message, i've changed the last series to be -ve, so this would suggest that this hypothetical strategy doesn't perform well when the book% is >110%?
If I am correct, this looks like something I would love to do more regularly. To throw in some ideas, instead of Book%, could bucket: market volume, race course, selection volume %, a custom indicator etc? I hope I have not missed the mark here and that it sparks some ideas for others.
As all of the results have a +ve expectancy, the graph shows all lines rising nicely. Is this how you would use this? Any series that goes -ve is a weakness of that strategy? Just before sending this message, i've changed the last series to be -ve, so this would suggest that this hypothetical strategy doesn't perform well when the book% is >110%?
If I am correct, this looks like something I would love to do more regularly. To throw in some ideas, instead of Book%, could bucket: market volume, race course, selection volume %, a custom indicator etc? I hope I have not missed the mark here and that it sparks some ideas for others.
You do not have the required permissions to view the files attached to this post.
quite possibly!
I think I might have to start not relying on backtesting so much, it's far too easy to just "fit" things it would seem...
-
Realrocknrolla
- Posts: 1903
- Joined: Fri Jun 05, 2020 7:15 pm
Starting to come to the conclusion it's just not possible to find an edge based purely on technical price action.
Which maybe the case, as with 1000s of very smart algo traders out there and throw in some advanced AI/ML systems, anything technical would be found very quickly and squashed...
Which maybe the case, as with 1000s of very smart algo traders out there and throw in some advanced AI/ML systems, anything technical would be found very quickly and squashed...
First, i've not got any inplay ideas as such! but thought i'd pursue the many many hours i've put into what i've done so far pre-race, before giving up on it.Realrocknrolla wrote: ↑Wed Sep 29, 2021 2:43 pmWhat happened to your Inplay ideas?
-
beermonsterman
- Posts: 522
- Joined: Sun Dec 25, 2016 2:47 pm
- Location: Birmingham UK
My automation does well and its a simple but affective bot they are out there goat just need to keep things simple
You do not have the required permissions to view the files attached to this post.