Number Crunching After Capturing and Sorting.

A place to discuss anything.
Post Reply
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

ruthlessimon wrote:
Wed Sep 19, 2018 9:28 pm
There will be a correlation between course volume @ 5mins & the average price of the fav @ 5mins
Probably yes, but that's mathematically inevitable. Rather than look at traded volume, look at traded liability (aka implied volume) and I'd guess it correlates with race quality more closely than price. It's simply not logical to compare the vol on a 1.5 horse to a 5.0.
User avatar
Euler
Posts: 24700
Joined: Wed Nov 10, 2010 1:39 pm
Location: Bet Angel HQ

Volume is dictated by many things but ultimately the price of the favourite is a key factor. The smaller the price the more money it will attract.
User avatar
jimibt
Posts: 3641
Joined: Mon Nov 30, 2015 6:42 pm
Location: Narnia

ShaunWhite wrote:
Wed Sep 19, 2018 9:41 pm
ruthlessimon wrote:
Wed Sep 19, 2018 9:28 pm
There will be a correlation between course volume @ 5mins & the average price of the fav @ 5mins
Probably yes, but that's mathematically inevitable. Rather than look at traded volume, look at traded liability (aka implied volume) and I'd guess it correlates with race quality more closely than price. It's simply not logical to compare the vol on a 1.5 horse to a 5.0.
when i went down this route (which simon will recall from 15-18 months ago), it was an interesting time. the combo of vwap/ltp seemed to work REALLY well. UNTIL - we had a scatty runner that threw the rider or a crisp packet blew across the stalls! (strange but true). Due to this, I now have doubts that you can trade via signals (pre race) without a major caveat. IF (big IF), one were able to seperately detect a huge odds spike (due to crisp bags etc), and smoothe it out (based on current liability), then maybe you would be able to crunch beyond mishap.

at present, i'm of the opinion that unforeseen elements are just too difficult to deal with. this almost squares with an old observation that the best risk/reward ratio lies outside the *golden* 5 minute pre window.

disclaimer - i no longer trade pre on automation (at present) due to the scenarios that can occur out of kilter pre race...

[edit] - agree that there's a golden MEAN when we reach anything below 2.0 odds !! (tho - can go one way or the other)... Fractals anyone??
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

ShaunWhite wrote:
Wed Sep 19, 2018 9:21 pm
I guess you didn't read my pm where I mentioned the idea of implied volume (prob not its real name, I made it up). That, or similar will remove that distortion.
Yeah I musta missed that ;)
ShaunWhite wrote:
Wed Sep 19, 2018 9:21 pm
I don't think low volume means lack of conviction. I think it just means there's so much racing that your average hard up punter can't afford to be spunking money on 150 races a week and picks what's decent to be interested in.
I'd be interested to know if that's a hunch, or a conclusion to a test you've already performed
ShaunWhite wrote:
Wed Sep 19, 2018 9:21 pm
You seem to be posting a lot of positive edge charts lately. Are you sure they're all take strategies? I wouldn't trust anything in excel that needed offers.
I could easily add a "handicap" to any strategy - in order to underestimate the profitability (probably worth doing i.e. -2 ticks). But my aim is to discover "general concepts" rather than tick perfect edges.

i.e. Is race type useful, - to which the above graph proves yes. (haps vs nov stks).

But also seeing concepts perfectly mirrored i.e.

Image

Clearly even if I miss a tick entering & exiting, there's a bias which is linking novices & mdns
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

ShaunWhite wrote:
Wed Sep 19, 2018 9:31 pm
Volume

It's at the mercy of....
This is my whole overarching point. Crossing that barrier to profitability is tiny. But becoming a "big trader" means we have to answer those questions (i.e. volume varies during the yr, the volume is different on Sat)

.. which is why Peter/most top traders adjust their strats during a Saturday to account for the volume "spurts"

But also that's a refinement that could easily be tested (how does this strategy perform @ Saturdays vs Weekdays)

-Basically.. what I'm tryin to say. It's vital we don't jump to conclusions when asking the hard questions. & these are bloody hard questions! (i.e. does love volume promote mean-reverting activity)
User avatar
ShaunWhite
Posts: 9731
Joined: Sat Sep 03, 2016 3:42 am

ruthlessimon wrote:
Wed Sep 19, 2018 10:00 pm
ShaunWhite wrote:
Wed Sep 19, 2018 9:21 pm
I don't think low volume means lack of conviction. I think it just means there's so much racing that your average hard up punter can't afford to be spunking money on 150 races a week and picks what's decent to be interested in.
I'd be interested to know if that's a hunch, or a conclusion to a test you've already performed
I don't need to test it, the bigger the event the bigger the money, fact.

yes there might be one or two races where interest has peaked and something different might happen, but I'm talking about general concepts. I'm not really into the idea of staring at my screen all day for the next 20yrs looking for edge cases.
ruthlessimon wrote:
Wed Sep 19, 2018 10:00 pm
I could easily add a "handicap" to any strategy - in order to underestimate the profitability (probably worth doing i.e. -2 ticks).
How will you work that out? The spread varies according to time, price and volume. And if you offer, then the fill rate is subject to the same factors.

Remember too that the biggest edge you're likely to find will be smaller than the margin for error in your theoretical tests.
ruthlessimon wrote:
Wed Sep 19, 2018 10:00 pm
But my aim is to discover "general concepts" rather than tick perfect edges.
wow

My decision loop is:
1. Find a pattern
2. Does it make money
3. Does it improve anything I've already explored
4. F'ck me the rent is due, Goto 1

Are you saying yours is :
1. Find a pattern
2. Goto 1

Who's funding all this purely academic research !
User avatar
Naffman
Posts: 5626
Joined: Sun Aug 11, 2013 5:46 am

You've been going 2.5 years now; from when you first started these 'general concepts' have mostly likely evolved with the market itself

I think you might place to much emphasis on stats and figures when most traders (even automated ones) trade on intuition
spreadbetting
Posts: 3140
Joined: Sun Jan 31, 2010 8:06 pm

Two and a half years is a long time especially with Simon's workrate, surely you must have come up with some strategies to stick thru the market by now? You should put a few hundred aside and manually trade what you think your 'bot' should do, you'll learn a lot more about how markets react to liquidity etc by watching them live rather than old data in excel.
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

I know there's plenty of people in my boat :)

Yeah absolutely, my issue is (& it's has always been) refining those strategies & continuing my growth.

I'm at a very awkward point, I know I've got half decent ideas/edges (not backing by letters!!!) - I just find refining those ideas, into an "all-encompassing, varied staking, approach" very very difficult. & I don't really want to post my trade ideas, but at the same time I don't understand my trade ideas :roll:

I remember back when I first started literally just backing certain faves. Almost random but slightly +ve. Technically, I had an edge straight off the bat, but I didn't want to do that for my entire trading career :lol:

This is a good analogy, which sums me up to a T:

I was at a really good footballing academy as a youngster (i.e. what Jimibt talks about above), but now I'm at this sodding Vanarama team, & I refuse to play - citing my academy pedigree!!! (but at least I'm at a team)

i.e. I refuse to use the edge because it's not good enough :roll: :lol:

That's such an awful mentality, but does sum me up quite well! I just don't want to be at a Vanarama club my entire career!!
spreadbetting
Posts: 3140
Joined: Sun Jan 31, 2010 8:06 pm

ruthlessimon wrote:
Thu Sep 20, 2018 5:13 pm
i.e. I refuse to use the edge because it's not good enough :roll: :lol:
Until you actually put money through the markets you'll never know if you have an edge in the first place. I don't know anyone who goes straight into the markets with a fully formed bot, you'll forever be tweaking them as long as you use them. What appears to work on paper doesn't neccessarily work in practice because the markets are so heavily botted your input into the market can create a butterfly affect even from relatively small stakes.
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

ShaunWhite wrote:
Thu Sep 20, 2018 2:22 am
I don't need to test it, the bigger the event the bigger the money, fact.
You've misinterpreted my point

At certain points during the market, I don't think volume matches what it should - & this concept would apply to both Lingers & Chelt (problem is that it's incredibly difficult to test via. data)
ShaunWhite wrote:
Thu Sep 20, 2018 2:22 am
How will you work that out? The spread varies according to time, price and volume. And if you offer, then the fill rate is subject to the same factors.
I'll send you an idea for you to test on your sheets - & you tell me if it isn't an edge according to your sheets ;)
ShaunWhite wrote:
Thu Sep 20, 2018 2:22 am
Who's funding all this purely academic research !
The irony is, that statement is pretty much true - hence proving why government agencies generally never get stuff done :lol: Although still.. I would've reworded it like this:

1. Find a pattern
2. Does it pay the rent & some, if no, step 1 :lol:

Anyone can have an edge just boarding on random imo!
User avatar
Naffman
Posts: 5626
Joined: Sun Aug 11, 2013 5:46 am

You're a perfectionist Simon ;)

There's so many right ways (and so many wrong!) to profit at trading that it's probably more of a hindrance to set your sights on one generic system
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

Naffman wrote:
Thu Sep 20, 2018 6:02 pm
You're a perfectionist Simon ;)
I must say, looking back, trading is probably the worst profession for someone with that trait.

I've gotta lose it, or it'll lose me ;)
User avatar
ruthlessimon
Posts: 2094
Joined: Wed Mar 23, 2016 3:54 pm

ShaunWhite wrote:
Thu Sep 20, 2018 2:22 am
How will you work that out? The spread varies according to time, price and volume. And if you offer, then the fill rate is subject to the same factors.
ruthlessimon wrote:
Thu Sep 20, 2018 5:43 pm
I'll send you an idea for you to test on your sheets - & you tell me if it isn't an edge according to your sheets ;)
Without going into the actual details of the strategy - yea or nay? Maybe post up your equity curve

& can you see why I call that a "Vanarama league edge" - clearly that can be improved with the right approach.

Moreover, the inputs of that strategy highlight some very key concepts imo - again concepts which I don't fully understand
Jukebox
Posts: 1576
Joined: Thu Sep 06, 2012 8:07 pm

Or possibly that you might not have chosen the right market for such perfectionism - pre off horseracing has so many variables and the last 10 minutes or so introduces so many more and so much more frequently that can suddenly change everything.
Whereas 6 dog racing with no jockeys, negligible loading issues, negligible delays or conditions changes to worry about or prematch soccer or tennis might be better suited to your approach.
Post Reply

Return to “General discussion”