50 A Day System

[rayza1978]

Has anyone tried Level Stakes and use in Handicap Races only??

I have not got enough data yet but looking back over the last 3 months looks profitable

Using the staking plan advised is pure suicide!

Ive used the Racing Post,and looked for the track with the poorest return for favs purely on track statistics on Handicap racing only and there maybe something here as the profits have been very good over the past 3 months but lots more data to collect

[Atho55]

This is based on Data starting on 1st Feb this year, any race with Hcap in the title, Rank 1 in order of BSP favouritism, £25 Lay bet AND only showing the Courses that have a +return assuming all bets placed @ BSP. Commission has NOT been factored in.

Lay Rank 1 Hcap Only.jpg

The screenshot is the biggest I could grab to show all the courses but gives you the gist. The chart is a cumulative from the daily returns starting 1st Feb. Other numbers count the days it has either won or lost.

[spreadbetting]

Rank 1 in order of BSP favouritism, £25 Lay bet AND only showing the Courses that have a +return assuming all bets placed @ BSP. Commission has NOT been factored in.

It's always easy to find positive trends by drilling down your data, I think the term is backfitting, the trick is to do it beforehand. You can't lay unnamed favourites at BSP and runners can flip flop numerous times before the off between many runners, plus why choose Feb and not March or the start of the flat etc

This week laying the top rated horse in handicaps/stakes races has acheieved a 100% profit, could be worth lumping on?

[ruthlessimon]

It's always easy to find positive trends by drilling down your data, I think the term is backfitting, the trick is to do it beforehand.

For the sake of being devils advocate - what if that genuinely was a "new edge"?

I just wanna caveat that by saying - no way would i trade it - because that is such a dangerous use of course imo (i.e. no obvious links between em/size)

But is it possible to pre-empt "new" technical edges without some form of backfit?

[spreadbetting]

No one's saying you should ditch your data but simply looking thru data to find a longterm 'edge' is like looking for a needle in a haystack as you'll always be able to drill down the data to find something that returns a positive outcome.

Back in the day when I first starting betting you'd find so many systems being sold for the weekends races with 80% + strike rates, they'd simply pick the day's big race, look thru the results and find some common theme that would have made a profit. Maybe backing the top weight had returned a profit if backing every running of the race and it be sold in the racing papers for a tenner or so. Once in a while the people buying those systems might have struck lucky when the topweight won that race.

I see nothing wrong with using data to try and confirm some theory but simply drilling thru data to find an 'edge' seems no different to the newspaper system sellers to me.

[rik]

This is based on Data starting on 1st Feb this year, any race with Hcap in the title, Rank 1 in order of BSP favouritism, £25 Lay bet AND only showing the Courses that have a +return assuming all bets placed @ BSP. Commission has NOT been factored in.


Lay Rank 1 Hcap Only.jpg


The screenshot is the biggest I could grab to show all the courses but gives you the gist. The chart is a cumulative from the daily returns starting 1st Feb. Other numbers count the days it has either won or lost.

you ran that in test mode or real?
i understand right you just put £25 lay keep bet at odds 2 on the BSP favourite?
Im pretty sure most uk race courses would behave quite similar when it comes to this, difference between race courses likely mostly variance.
race distance could make more difference. Irish race courses i believe there must be more time advantage of the faster guys as there are bigger price jumps towards the end of the race which could have some bets matched at very bad value.
whats the win/loss ratio, ROI?

[ruthlessimon]

Back in the day when I first starting betting you'd find so many systems being sold for the weekends races with 80% + strike rates, they'd simply pick the day's big race, look thru the results and find some common theme that would have made a profit. Maybe backing the top weight had returned a profit if backing every running of the race and it be sold in the racing papers for a tenner or so. Once in a while the people buying those systems might have struck lucky when the topweight won that race.

I see nothing wrong with using data to try and confirm some theory but simply drilling thru data to find an 'edge' seems no different to the newspaper system sellers to me.

Ty for that I guess at the end of the day it comes down to humility & rather big dose of common sense.

I defo agree that Atho's use of data is reckless - that use of course makes me cringe (I'd be interested to know if it was still profitable without that caveat). that said I do think certain discrete variables can be useful, but they have to have some core logic imo (i.e. uneven distribution, similar characteristics, pure outsample) - albeit usually I'll "make up" ( ) the logic after I've found something

[Atho55]

I started collecting data on the 1st Feb so can only go back that far. The OP had noted a trend and asked specific questions, I gave him some answers it`s as simple as that. No backfitting just as it is.

Within BA I believe you can use the condition Actual SP = Actual SP of Position 1 (sorted in order of Favouritism) to place your bet but you would need to be a subscriber to know that.

If you have more data, extend the range for another year and post it up using the same courses and BSP Rank 1

I`m not trying to find an edge, I`m answering the OP.

[spreadbetting]

................. that said I do think certain discrete variables can be useful, but they have to have some core logic imo (i.e. uneven distribution, similar characteristics, pure outsample) - albeit usually I'll "make up" ( ) the logic after I've found something

The problem with most of those 'discrete variables' is that they're not acheivable in the real world giving false results that look good on paper but rarely translate well in the markets.

Within BA I believe you can use the condition Actual SP = Actual SP of Position 1 (sorted in order of Favouritism) to place your bet but you would need to be a subscriber to know that.

You'd also need to know the exact off time the race will go off to accurately predict the horse with the lowest BSP and I doubt BA knows that.

[Atho55]

Using SV` s at 1sec into the race you can find the Rank, then you place your bet....

BSP to Rank.jpg

I capture mine on the OCS but to trigger a bet it works reasonably well. You may not always get matched at BSP but that is another issue. The data uses BSP.

Are you doing another year by any chance as you appear to have the data?

[spreadbetting]

Are you doing another year by any chance as you appear to have the data?

Here's your Handicap data from the Betfair data files, in CSV data but had to zip it because BA doesn't accept CSV for some reason.

Here was the query if there are any obvious errors as I'm a MYSQL novice

SELECT `MENU_HINT`,`EVENT_NAME`,`EVENT_DT`,min(`BSP`),`WIN_LOSE` FROM `new_table` WHERE `EVENT_NAME` not like "%Placed" and (`EVENT_NAME` like "%hcap%" OR `EVENT_NAME` like "%nursery%") group by `EVENT_ID`

new_table(1).zip

Edit , tweaked the file as I'd forgotten to include Nursery's

[ruthlessimon]

Interestingly it passed my reverse out of sample.. however, when we apply it to all courses it's unprofitable

Also commission seems to have a huge impact with these types of trades

0% -£3k
5% -£15k

[Atho55]

When I said are you going to do the earlier year I meant actually do it to show the Pivot Table and chart, not send the data for me to do another.

Anyway had a look, how do we find Rank 1 as that is part of the brief OR are the records you sent ALL the Rank1`s

If these are NOT the Rank 1`s how have you managed to do it Si

If they are the Rank 1`s, do the Pivot table and Chart as well Si

I believe I mentioned that commission was not considered

[spreadbetting]

I'd have no clue as how to do a pivot table

It's just a csv file of the lowest BSP runner's odds in every handicap since around 2008 from the betfair data files, contains MENU_HINT`,`EVENT_NAME`,`EVENT_DT`,min(`BSP`),`WIN_LOSE` which are respectively, the Race, the race details i.e. distance etc, the date/time , BSP (which was the lowest in that race) and a flag in the win_lose column that represents if that lowest priced runner won (1,2=dead heat) or lost (0)

[ShaunWhite]

This is based on Data starting on 1st Feb this year, any race with Hcap in the title, Rank 1 in order of BSP favouritism, £25 Lay bet AND only showing the Courses that have a +return assuming all bets placed @ BSP. Commission has NOT been factored in.

I know this isn't your battle to fight Athos55, I'm just quoting this as a good example of how misleading data can be.

These were the courses with a +ve return last month, how does that have any bearing on the courses with a +ve return next month? (rhetorical question)

Trends have a dangerous psychological effect on some people but my instinct is always to opposite them; if there's large scale efficiency then mean reversion should be respected.

Take the Ashes this year, the first post on the thread said ...

Australia haven’t won an away Ashes series since 2001 and start at Edgbaston where they haven’t won in any format since 2001 also, and only won 5 of their last 17 away tests. 8 of the 9 Ashes series have been won by the home side.

It's hard to see a series draw unless weather becomes a major factor.

.....I know absolutely nothing about cricket but my 1st instinct was to think "Well, England probably won't win then, those are streaks that can't go on much longer". And they didn't (I don't think). I'll happily oppose any noteworthy streak or trend, unless there's a compelling reason not to. The only reason I'd look for courses where the fav was value would be to oppose it.