The home of intelligent horse racing discussion
The home of intelligent horse racing discussion

Statistics

Home Forums Archive Topics Statistics

Viewing 15 posts - 1 through 15 (of 17 total)
  • Author
    Posts
  • #4042
    rory
    Participant
    • Total Posts 2685

    Possibly slightly beside the point, but when studying stats at school, we constantly talked about "statistical significance";  the concept being that a correlation was only held to be solid if it occurred in more than a certain % of cases.  Do any published statistics go through such a process of verification?  Could this be done by the layman?  If so how, and using what information?<br>

    #93878
    Sal
    Member
    • Total Posts 562

    Statistics on their own can be used to make an argument from any angle. I suppose the thing to always remember is the events that are behind the statistics. The figures themselves should just be a useful way of summing things up.  For (a very simplified) example – a sire’s figures tell you that 97% of his progeny do not like soft ground.  You know the horse you want to back is one of the 3% that has already won on soft. In this case the sire statistics, although overwhelmingly against your selection, are not backed up by the evidence of this event.

    However, the racing statistics of the individual horse involved tell you that the horse has won on the soft in the past.  This would suggest that he is quite likely to win in it in the future.

    I guess what I’m trying to say is that statistics can be useful – as long as you look at the right ones!  It should’t ever be a matter of blindly following numbers.

    #93880
    LUKE
    Member
    • Total Posts 271

    The one statistic I always relied on heavily was that 2 out of 5 odds on shots gets beaten.

    #93884
    Steve M
    Member
    • Total Posts 99

    Statistics are of tremendous importance, the trick is in having the sense to realise which ones are meaninful and which are not.

    #93889
    robgomm
    Member
    • Total Posts 224

    Hi Razeen,

    I think there’s a reason for favourites having bad record in some races: There are many big, competitive races about where favourites have bad records. Like the recent Gold Cup’s. Plenty of favourites have lost the Gold Cup in the last 10 years – i think only one has won the Gold Cup, Master Oats. Punters will back the favourites in competetive races (like the Gold Cup) and the horse’s price will shorten. However, the punters seem to get it wrong a lot of times with favourites in big, competitve races.

    Favourite backers are going to back the favourite in big races rather than go for a longer priced outsider. The favourite shortens. A longer priced horse comes in because the punters have got the wrong one.

    Sometimes the favourites win if the punters get it right but in the competitive popular betting heats it is possible that many favourites are bad value.

    (i don’t have any statistics to back this up though…just a simple theory that many have thought of)

    #93890
    pixo
    Member
    • Total Posts 32

    99% of the time Data Statistics is the only input that determines weather I have a bet and its stake (Im the cold-analytical type of gambler I guess)

    (Edited by pixo at 9:01 pm on Jan. 29, 2002)

    #93891
    Steve M
    Member
    • Total Posts 99

    If you have a statistic that no blue colt has won the Derby since records began, it is a meaningless statistic (as there are no blue colts, or at least none that have run in the Derby).

    If you have another that fillies are not prone to winning the Derby, it is virtually meaningless as so few run in it. Therefore you would have to perform an analysis of whether in percentage terms fillies were any less likely to win the Derby than they are any other all sex race.

    However, if you have another statistic which tells you that it is virtually impossible for a 5-y-o to win the Stayers Hurdle, you ought to take note, as many run in it and they invariably fail. So there are likely to be very good reasons why they don’t win.

    This is what I mean Razeen.<br>

    #93892
    pixo
    Member
    • Total Posts 32

    Razeen, a simple profitable statistic off the top of my head would be, for example, the record of favourites (Rob take note) in the St Leger but, more often than not, profitable statistical data may contain more variables, be rather complicated and time consuming to unravel.

    #93893
    jjimps
    Member
    • Total Posts 43

    The thing about statistics is they are rally only usefull once you understand the reason for them. Just blindy backing on stats alone is a road to nowhere but if you can uncover the underlying reason that has lead to the figures you are on a winning trail.

    I will illustrate with a question :<br>Imagine a race track where<br>1) There is a 5f straight course on which the draw produces no advantage<br>2) there are 10 stalls<br>3) races are run fairly and the winner wins on merit every time

    When I study the results of 100 races run over this course what do I see with regard to winners from each of the 10 draw positions ?

    #93894
    Steve M
    Member
    • Total Posts 99

    No it is not something that is factored in by bookmakers as 5-y-os can start the race as favourite or near favourite, when in effect they have no chance. In other words it is worth knowing that 5-y-os just don’t win.

    And yes you can tell if it is statistically significant without looking at the whole population. It becomes statistically significant for that sample. The bigger the sample the better sure, but the sample for runners in the Stayers Hurdle is large.

    The reasons 5-y-os don’t win the Stayers is I suspect to do with them not being strong enough in relation to those that are older, even when they have shown great promise in their own age group or at shorter distances against older horses.

    alsoran, what you say is true. The sophistication needed to improve your system would be one that gives weighting to the parameters.<br>

    #93895
    tooting
    Member
    • Total Posts 379

    Razeen, that’s a good point you make higher up this thread, about whether ladbrokes are factoring your statistic into their prices.

    It also applies to guessing whether the betting public are taking any one statistic into account.  We do collectively tend to over-exaggerate the effect of certain factors in certain circumstances.

    If there’s four or five items in the racing press suggesting (for example) 5 year olds can’t win the Stayers Hurdle, not only may ladbrokes factor that in as a negative in their pricing of a decent 5 year old, but the horse may drift even further because everyone’s cottoned onto this statistic, and then probably drift further still once on the slide as the rest of the public think there must be something wrong with the horse.

    At some point the 5 year old is going to become seriously good value, regardless of the statistic.  

    I use quite a lot statistics in forming my books (trainer form in particular). But I’m always on the look-out for the sort of statistic that is going to become a monster and affect the book out of all proportion to its real meaning.   <br>

    #93896
    Seagull
    Member
    • Total Posts 1708

    Lies, Dam Lies and statitistics that the old saying but I know in racing they work.<br>On average 33.3 % of favs will win so whatever the price of the last one of the day if 5 have won (on a 6 race card) the % of the last fave winning is under2% yet if none have won it is just over 98% . That could be useful for example.Tke Nods Nephew a week agoan example of that stat in action!<br>The best are trainers stats it took Epsom tainer T.Mills seven seasons to have a winner at Epsom.<br>Even if you back every Godolfin runner the last 3 seasons you would have made an overall profit .<br>Mary Reverley an exellent trainer but the stats show terrible results at Cheltenham and yet another joint fav at the last meeting (got well beat) I layed it in Betfair.<br>Of course some stats are useless but there are ways of using them to your advantage.

    #93897
    Seagull
    Member
    • Total Posts 1708

    I mean that although the stats may not help you find a  winner Mary Reverley at Cheltenham for example (terrible strike rate for whatever reason) BUT you could use this knowledge to oppose her and lay them in Betfair.Some are just so obscure they are not worth botherering about how many horses win after 2 runs or 3 runs for example. Believe me it is all available!<br>try this site http://www.racedata.demon.co.uk it is exellent

    #93898
    rory
    Participant
    • Total Posts 2685

    Seagull

    I take it your favourites stat was meant in jest?

    #93899
    Merlin the Magician
    Member
    • Total Posts 152

    NATIONAL-HUNT RACING, I myself do not believe in stats for stats sake people on here go on about certain races ran on the same day I.E. the novices time was 2 seconds faster than the grade 2 race that just took place? dont ever get confused with this type of stat each race MUST be judged on its merits  its MORE TACTICAL in any one race than a thing with times, but of course some stats are of tremendous importance,C-D is one that a punter must give a lot of time too, GOING-DISTANCE another but all the other stats are just to keep someone in a job totally irrevelant this is based on my way of gambling NATIONAL HUNT ONLY <br>Regards MERLIN

Viewing 15 posts - 1 through 15 (of 17 total)
  • You must be logged in to reply to this topic.