Differences in advanced_going and going

0 votes
Apologies for the bombardment of questions, but I've just noticed something else that is slightly skewing my historic picks against my dailies.

It has to do with the advanced_going field in historic_races_beta and the going field in daily_races_beta.  It appears that on some days, there are some differences between these fields.  I appreciate we can't always know what the daily going is going to be exactly at 7.30pm the night before, but it would be useful if the two matched.

I've noticed a few other posts about the same thing so I was wondering where we get the historic_races_beta 'going' data from?  It appears to be simplistic compared to what's in the daily_races_beta table.  Would it not be a good idea to provide updates to the daily_races_beta table just before the races start, say midday, that will match what the 'going' data will be in the historic table which is added a few hours later?

It seems like a fairly straight forward process that would definitely help in my situation (and others according to the other similar questions).  At the moment, I can do historical analysis but it's dependent on the 'going' field.  When I come to do my daily picks, quite often the picks are different to what will be chosen when I do the historical analysis for that day (the following day).

asked Sep 27, 2023 in Smartform by maddisor Novice (410 points)

1 Answer

0 votes
Unfortunately, daily going descriptions are constantly changing in advance off racing and even during racing, and the descriptions are highly heterogenous in nature, so there is no easy solution, however we provide a number of utilities to assist with a) keeping on top of changes in going and establishing what the going is, and b) comparing going.

Firstly, all subscribers have access to the race_going pages from the members area, and can ingest these changes - along with corresponding race_ids matched to Smartform, as part of their daily updates - we would ask that updating is done in fair use terms (eg. at midday and possibly just before the race off time).  

Secondly, we provide simplified going descriptions to fit in with the purpose of what you are trying to achieve above within the *insights tables.   See the simplified_going and grouped_going fields (and corresponding descriptions) at:



Last but not least, users can apply their own transformations to the daily going data, using the type of logic described above, since descriptions are often highly variable and open to interpretation.  For this very reason, there is ongoing work in British racing to standardise going descriptions against quantifiable criteria, as well as to measure going outside text based descriptions issued by the Clerk of each course.
answered Sep 27, 2023 by colin Frankel (19,320 points)
edited Sep 27, 2023 by colin
Thanks Colin.  I did investigate using the insights tables, but the simplified_going field is different between historic_runners_insights and daily_runners_insights, at least one course a day is completely different.  e.g.

select distinct(ins.meeting_date), ins.course, ins.simplified_going, da.simplified_going from historic_runners_insights ins
inner join daily_runners_insights da on da.race_id = ins.race_id
where da.simplified_going <> ins.simplified_going

I was expecting this to return barely any rows, but I got 1071.

I couldn't find the 'race_going' pages in the members area - would you mind forwarding a link please?

I guess one option for me is to run my daily picks using goings either side of those which appear on the likes of RacingPost and SportingLife in the hope I'll catch any daily picks my historical analysis picks up.
The historic and daily insights tables will change due to going differences based on weather changes from advance racecards to time of race - the descriptions are however homogenised.
I've run my historic analysis over 1000 days using the simplified going and I'm only slightly down in terms of win ratio from my previous goings scale.  Problem is, I'll need to try and match the daily goings with those 5 simplified ones.  It's a shame they don't match between historic and daily insights.
They can never match exactly, due to real differences and the fact that the daily record is always an advance record, which we do not alter.
Course stewards are often reluctant to report a change of course going in advance, but the true assessed going will be reported with race results. That could in principal differ for every race if there are weather changes or different course sections are used. Base your ML training and back testing on results reported going. You can sometimes use results of prior races or sectional times, where anticipated performance differs from expected performance based on going, to "improve" your assessment of actual current course going. I have seen cases where I know it's raining, runners are strongly suggesting the course has gone heavy, but the stewards are not reporting the going change, possibly for fear of getting too many non-runners.