Chrome Prince,

A Horse that has won 4 times at Caufield from 4 starts would have strong claims to my bet [ amongst other considerations including value ] even though it might start favorite. Good form at a track usualy means more good form to be expected [ but not always naturaly ]. So what do we do , dump horses with good stats. because they will shorten the price beyond value ? Value is very subjective and the ability to disern it and measure it is owned by no single being. How many punters who visit this forum backed FOO racing in Hong Kong at what was it $3.50 somthing here when it was $30 over there ? Obviously to some it was a value bet.

'Horses for courses' has long ago become a figure of speech that can mean many things, even about things that have nothing to do with horses : 'sharp knife for a steak, not chopsticks' [ horses for courses ]. I used it to mean an orderly approach to future punting as in worry about one race at a time.

My point regarding backfitting was just a light hearted prod at the use of Stats. to bludgen a point to death. The problem of using stats to deal with selection or staking is that apart from being a great selection aid they can also lead to assumed wisdom and rightiousness, not just valid speculation amongst other valid speculation.

Cheers.


[ This Message was edited by: crash on 2003-12-10 23:41 ]

Crash,

See your point.

In fact on this very forum I said that FOO would run close to last in Japan based on my original argument re the time run in the Cox Plate.
I said I'd be laying FOO on Betfair for as much as I could - and I did.
:smile:

Re stats, no righteousnous, but wisdom can come from a decent sample of stats.
The problem is that some draw conclusions from even 1,000 races and this is not nearly enough.
I have recently expanded my database to over 14,000 metro class races in order to identify the trends which repeat year in year out - these trends cannot be ignored.

All the best.

Gee Chrome don't start sounding like a politician, they are all wisdom when it comes to Stats. Just listen to those little humbug sprukers from Camberra!! You are right though about the need for very large amounts of stats to identify trends and in that role they are great. Not bad for handicapping purposes either.

FOO was looking forward to a very large paddock not another race.

Cheers.

Crash,

I see a huge difference between analysing stats to see if they refute or support a particular argument and backfitting (which in my book means applying different rules and filters to data until you come up with something that shows a profit, regardless of whether it is logical or not).

Your original comment on this thread was:

Anyone backing outside barriers hasn't been punting long enough to know better and I don't care if the stats. show it's almost a level playing field.

To me this makes no sense - you are saying in the one sentence that you think outside barriers are bad but acknowledge that the statistics show there is no difference! Why are you opposed to outside barriers if you have never bothered to check what the results would have been from including them in your betting???

The statistics I quoted showed that barriers do have a small impact on the strike rate but that punters obviously over-estimate this affect as average dividend goes up by more than it should given the small disadvantage.

I guess my point is that at least I am basing my decision to generally ignore barriers on some factual evidence rather than just a perception that they are bad. The strange thing about human perception is that two people can view the same events and come to totally opposite conclusions depending on their previous assumptions or beliefs.

There have been some really interesting psychology experiments that I have read about that basically show humans will remember things that confirm their beliefs but dismiss and forget those that challenge the beliefs - therefore over time your beliefs are reinforced even when they are not necessarily true. In one experiment half the subjects were given a very convincing (but totally false) lecture on why left-handed people are worse at catching balls than right-handed people (or it might have been the other way round). They then watched a video of a baseball game with a team of "righties" playing a team of "lefties" and at the end had to estimate how many catches each team had dropped. Something like 90% of the subjects who had received the lecture over-estimated the dropped catches by the left-handers and underestimated the right-handers whilst the ones that did not receive the lecture were much better at estimating the true numbers (which were, in fact, the same for both teams).

With the barrier example each time you see a race where the favoured runner starts from a wide barrier and is beaten your brain will see this and store it away as confirmation that your previous perception of barriers is correct. When the winner starts from an outside barrier you will more than likely not remember the barrier position or dismiss it as it conflicts with your perception. It is only when you see the real figures down on paper that you can make a judgement on whether your perception is true or false!

PS. If you are going to put my home town down at least spell it right! Of course Canberra actually has the lowest number of politicians in permanent residence of any Australian state or territory - all the w**kers live out there with you guys and only visit here for a few weeks at a time (and only during the week - you should see the suits at Canberra Airport on a Friday evening)!
_________________
"Computers can do that????" - Homer Simpson

[ This Message was edited by: becareful on 2003-12-11 13:19 ]

Hi crash,

Politicians use stats to hoodwink the public.
Did you know that the reason that the unemployment rate has dropped to around 5% is purely because JH takes people on the dole off it, and sends them to study basic courses at TAFE.
They are then transferred OFF unemployment and onto study allowance.
Thus, the stats say there are less unemployed technically than in reality.

It's akin to sending all criminals to New Zealand for a day, then doing a survey of criminals in Australia and after the survey's done, shipping them back here.

headline: "Australia has the lowest criminal record."

This is exactly how it's done.

This is false and misleading.

Enough politics.

:sad:

BC, just went thru the stats in the first pages.
Are you able to filter the same using slow/heavy/wet/raining?

Let's say a heavy track, majority of winners from which barrier numbers.

So JH started the fiddling of the unemployment figures.
I thought it started 30 years ago with the increase in invalid pensions etc. and disabilities, increased school leaving ages and all those things. statistics are statistics. but having said that i dont believe government is necessarily the party of the day. government is the departments who introduce policies for their current politicians to carry out.
its a pity there isnt a department for sydney racing then it might improve also.

Puntz,

I reran the 14+ runner scenario for Slow and Heavy tracks only. To get a more reasonable number of races I extended the timeframe back to the beginning of 2001 which gave about 500 races. The results were:

B1 to B7: 7.6% strike rate, $11.30 Ave Div, 14% Loss
B8 to B13: 6.2% strike rate, $11.21 Ave Div, 30% Loss
B14 to B24: 7.5% strike rate, $12.40 Ave Div, 6% Loss

Interestingly it is quite consistant with the all tracks data.

Looking at this maybe I SHOULD be considering barriers in larger fields - but the danger area is not the outside as most people seem to think but rather the MIDDLE barriers as these show a big dip in the strike rate!

_________________
"Computers can do that????" - Homer Simpson

[ This Message was edited by: becareful on 2003-12-12 08:11 ]

Those figures are pretty much what I would have expected. Any disadvantage that might exist from a wide barrier is more than compensated by increased dividends.


[ This Message was edited by: sportznut on 2003-12-12 09:40 ]

Thanks BC.
Place bets on the outsiders I suppose on heavy tracks.
Victoria Park brings back memories one day punting.It was pouring and listening on the radio how these mud-grubbers were coming in, using the "better" section of the track *from* about race 4 and onwards. No need for rocket science to work out why.
I know it POURED with rain and outside barriers won most of the day.