Well Chief, I think it's unfair to judge Finians Rainbow's Arkle against Sizing Europe's at this stage. For instance Somersby has only won his Grade 1 in recent months.

Captain Chris has had an unlucky season imo. He unseated when looking like the winner in the Haldon Gold Cup. Then in the King George he was running a cracker before blowing up. A line can be drawn through his last run. Finians Rainbow has ran as well as Somersby did in the same period between Arkle and Champion Chase last year. Medermit won the Haldon and ran a cracker behind Quantitativeeasing at Cheltenham. Realt Dubh has yet to run this season and could still be a serious horse. The very fact that they are willing to run him in the Queen Mother first time out tells you how highly he's regarded.

My point is that we've had a year extra to judge the form of Sizing Europe's Arkle, which before Somersby won a Grade 1, I wouldn't say looked very strong.

Thank you for all the feedback.

Have also managed to get some feedback on twitter from some very knowledgable people in the industry. Richard Hoiles and Simon Rowlands (head of research and product development at timeform).

Both raised the point that ratings horse achieve don't tend to have a normal distributon. Simon commented that they had a skewed distribution. He also took the time to email me with a more detailed response.

Below is the response I got from him if people are interested in the topic(I hope he doesn't mind me posting it on here!)

First off, I hope you did not mind me disagreeing with one of the premises of your piece. I did think it was very refreshing to see an intelligent and alternative approach to form analysis: well done. Much of what else is out there is tired and unimaginative.

Hopefully the attached will give a picture of ratings distributions. Whereas "master" (i.e. the one rating used as definitive) ratings do distribute close to normally over a horse population (though there are issues with opportunity presented by the race programme and a tendency to a long tail at high values), the array of "performance" ratings of a horse do not.

My example uses a smallish sample, simply as I had to extract them manually, but still illustrates the point: the graph is heavily skewed towards a horse "running to form". When you think about it, a horse's maximum performance rating has to exist at least once for each horse, and therefore the frequency with which it appears must be at least 1/N where N is the number of runs being considered.

In addition, we are not observing a natural phenomenon but one which is greatly prone to human intervention, notably in terms of a trainer/jockey trying to ensure a horse runs well (well, we hope so anyway!). A horse is not raced at random, on a random day, under random conditions, and without being trained to peak for racecourse performance. The distribution should skew towards running well.

I mentioned the possibility of conducting a Monte Carlo simulation with arrays of ratings. In simple terms, this might be something like:

Sizing Europe performance ratings: 172/172/160/146/164/172/157

Finian's Rainbow performance ratings: 166/153/159/155/138/142/142

If every performance rating was chosen at random countless times, and the higher/highest rating "won" the race (a MC-sim), you would get Sizing Europe "winning" ((3*100)+(1*86)+(1*86)+(1*72)+(1*42))/7 = 84% of the time and Finian's Rainbow winning 16% of the time.

However, you need to do this for every possibility in a race, which may involve many more than two runners and arrays covering more than seven performances. In addition, this works better with exposed horses with established form - such as handicappers - than horses that were and may still be progressing, like this pair.

Nonetheless, it is possible to write this all into a spreadsheet. I know for a fact that many professional punters who employ mathematical modelling use a degree of MC-sim in their conversion of ratings into probabilities (for both win and place).

Nobody want to use RT to **** SEs Arkle? He was in the race I believe. Somersby ran a cracker in th Supreme the year before as well. I just don't think he ttained as much as others.

I really wanna find something against SE in this but there looks to be nothing doing. I think it may be a no bet race or maybe just a place bet on Big Zeb.

theory goes to the test today extreme, personally haven't commented as didn't agree with it all and i see other decent judges on here did to a degree...all racing opinion i suppose...so will keep a lid on it as like i say didn't agree with it...

de bromhead has never made mistakes before and hope the waffle that i have been through doesn't bare out..

Great write up a pleasure to read. I am with Sizing already but may have a little interest e/w on Finians at 11/2 a free bet if ever there was one IMO. Good luck with your staking plan.

Good call EC,proven correct

suppose,was refering to the fact that many believed that there was no way FR could match the figure that SE ran to

a fact not mentioned ,but worth noting ,SE was ythe highest rated hurdler who went chasing for quite a few years ,if memory serves me.so this may have lead to higher ratings