![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
Today's Greyhounds
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
Too much data to play with. ![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
You do not have the required permissions to view the files attached to this post.
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
If you go to greyhoundstats website you can search for a dog and find it’s entire running history by grade, distance, trap, ruining time and sectional times.
- MemphisFlash
- Posts: 2226
- Joined: Fri May 16, 2014 10:12 pm
- Location: Leicester
you mean like this
You do not have the required permissions to view the files attached to this post.
Last edited by MemphisFlash on Fri Jun 23, 2023 11:25 pm, edited 1 time in total.
- MemphisFlash
- Posts: 2226
- Joined: Fri May 16, 2014 10:12 pm
- Location: Leicester
i have mine on one click screen
You do not have the required permissions to view the files attached to this post.
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
Jesus Memphis.
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
- MemphisFlash
- Posts: 2226
- Joined: Fri May 16, 2014 10:12 pm
- Location: Leicester
or these screensArchery1969 wrote: ↑Fri Jun 23, 2023 10:58 pmJesus Memphis.![]()
You do not have the required permissions to view the files attached to this post.
- MemphisFlash
- Posts: 2226
- Joined: Fri May 16, 2014 10:12 pm
- Location: Leicester
or even this
You do not have the required permissions to view the files attached to this post.
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
I’m actually looking into something which doesn’t care about any type of stats.
Got the idea from something Shaun mentioned.
- Wait for selections traded volume to be > £1000
- Store HTP and LTP
- Take the mid point
- Back anything above and/or lay anything below
Probably not exactly what Shaun is doing but in theory should give value and whatever the outcome EV+ over the long term.
Will see how it pans out and maybe build the VWAP into the mid point etc.
Doesn’t matter if it’s a dead end as live this kind of stuff.
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
I don't remember saying anything like that?
If anything it was far simpler. No check on traded vol (but do check for overround and a small spread), and I don't look at HTP or LTP., just the current offers. It's not really that groundbreaking, it's just using the principal of Efficient Market Hypothesis (ie all prices are always correct) and trying to get offers filled at mid + margin.
The problem with these theoretical mathematical things (incl random walk etc) is they only work at large scale because the edges are tiny. You'll need to be trading every dog/horse, probably over long periods and looking for several thousand small matches a day. If you don't then adverse selection wipes out any edge you might have. It's also harder to execute than it sounds as the margin you're asking for can change over time, as does the amount of time you leave your offer before moving it. I don't think it's possible to tune something like that without api (tick) data and a simulator, and when you execute it then the time between your offer and it being matched is a key metric for knowing what value you're getting (fast matches being poorer value).
It's an HFT strategy and as good as BA is, it's not designed to do that sort of thing. Take Guardian market cycling, if you load 300 markets you're refreshing every 6s at best. An HFT strategy needs to be responding within a few milliseconds on every market simultaneously. I don't cycle as BA does, I send each market change message to it's corresponding market as they arrive. I'm OK about describing it because I'm confident BA can't do it, and people not using BA will probably be familiar with that sort of strategy anyway.
![Smile :)](./images/smilies/icon_e_smile.gif)
If anything it was far simpler. No check on traded vol (but do check for overround and a small spread), and I don't look at HTP or LTP., just the current offers. It's not really that groundbreaking, it's just using the principal of Efficient Market Hypothesis (ie all prices are always correct) and trying to get offers filled at mid + margin.
The problem with these theoretical mathematical things (incl random walk etc) is they only work at large scale because the edges are tiny. You'll need to be trading every dog/horse, probably over long periods and looking for several thousand small matches a day. If you don't then adverse selection wipes out any edge you might have. It's also harder to execute than it sounds as the margin you're asking for can change over time, as does the amount of time you leave your offer before moving it. I don't think it's possible to tune something like that without api (tick) data and a simulator, and when you execute it then the time between your offer and it being matched is a key metric for knowing what value you're getting (fast matches being poorer value).
It's an HFT strategy and as good as BA is, it's not designed to do that sort of thing. Take Guardian market cycling, if you load 300 markets you're refreshing every 6s at best. An HFT strategy needs to be responding within a few milliseconds on every market simultaneously. I don't cycle as BA does, I send each market change message to it's corresponding market as they arrive. I'm OK about describing it because I'm confident BA can't do it, and people not using BA will probably be familiar with that sort of strategy anyway.
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
Just kidding Shaun. Its the heat and the Peroni.ShaunWhite wrote: ↑Sat Jun 24, 2023 11:37 amI don't remember saying anything like that?![]()
If anything it was far simpler. No check on traded vol (but do check for overround and a small spread), and I don't look at HTP or LTP., just the current offers. It's not really that groundbreaking, it's just using the principal of Efficient Market Hypothesis (ie all prices are always correct) and trying to get offers filled at mid + margin.
The problem with these theoretical mathematical things (incl random walk etc) is they only work at large scale because the edges are tiny. You'll need to be trading every dog/horse, probably over long periods and looking for several thousand small matches a day. If you don't then adverse selection wipes out any edge you might have. It's also harder to execute than it sounds as the margin you're asking for can change over time, as does the amount of time you leave your offer before moving it. I don't think it's possible to tune something like that without api (tick) data and a simulator, and when you execute it then the time between your offer and it being matched is a key metric for knowing what value you're getting (fast matches being poorer value).
It's an HFT strategy and as good as BA is, it's not designed to do that sort of thing. Take Guardian market cycling, if you load 300 markets you're refreshing every 6s at best. An HFT strategy needs to be responding within a few milliseconds on every market simultaneously. I don't cycle as BA does, I send each market change message to it's corresponding market as they arrive. I'm OK about describing it because I'm confident BA can't do it, and people not using BA will probably be familiar with that sort of strategy anyway.
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
You using VWAP ?
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
Kidding? I wrote all that for a kiddingArchery1969 wrote: ↑Sat Jun 24, 2023 12:54 pmJust kidding Shaun. Its the heat and the Peroni.![]()
You using VWAP ?
![Rolling Eyes :roll:](./images/smilies/icon_rolleyes.gif)
![Very Happy :D](./images/smilies/icon_e_biggrin.gif)
I don't use vwap, I looked at it but for me I couldn't get much out of it. It gets messy real quick. The std approach is to use 3 different duration vwaps (long, med, short) and look for them all pointing in the same direction but traded volumes are only indicative for the last few mins, and vary per market, even per selection, so it's just a headfck. There might well be something in it but I couldn't see it.
These last few months I've been experimenting with a variety of relatively simple approaches, eg changes in selection traded volume in last minute or so before the off, and also LTP changes (eg steamers) in the same or similar time period. And placing back bets accordingly.
The results have been variable but generally too -ve for my liking.
I have thought about using conditions to stop when either profit or loss for the day has reached a predefined level.
I haven't gone into the mechanics of each dogs history I'm afraid !
I have had previous experience as a runner, and was also inspired by https://youtu.be/sDZfLNcS-bc
The results have been variable but generally too -ve for my liking.
I have thought about using conditions to stop when either profit or loss for the day has reached a predefined level.
I haven't gone into the mechanics of each dogs history I'm afraid !
I have had previous experience as a runner, and was also inspired by https://youtu.be/sDZfLNcS-bc
- MemphisFlash
- Posts: 2226
- Joined: Fri May 16, 2014 10:12 pm
- Location: Leicester
Not A Fan Of Caan. He says a lot and says nothing at the same time
-
- Posts: 3314
- Joined: Thu Oct 24, 2019 8:25 am
- Location: Newport
Caan never explains anything but then people won’t as they not going to give away any type of edge.
Most of Caan’s trades look random entry to me.
Most of Caan’s trades look random entry to me.