Value odds

[Sparky12]

Hi All

Like many others, I use an Excel based ratings system to determine my picks for any given race (thanks to Shaun on this forum for technical assistance with this a few weeks ago).

Results so far are encouraging and my top 4 rated will produce the winner 68% of the time based on a sample of 200 races. POT of 21% backing all four horses to win to level stakes (using TAB divs as basis for calc). No doubt this'll reduce substantially when I eventually find the courage to put some cash on my selections!

My question relates to the concept of value. With a 68% chance of one of my Top 4 coming in the winner, value theory (as I understand it) suggests I should bet only when the combined odds on offer are below 68%? Is this correct? I understand that when backing a single horse, value can be achieved when my assessed odds for a specific horse are lower than the odds on offer from the TAB/Bookmaker but I wonder whether this approach is valid for backing four horses in a race when my assessment of their % chance of winning (68%) is based on a historic trend in a relatively small sample as opposed to taking each race and each horse on its merits and trying to establish value on a race by race basis.

Any thoughts on this? I've rerun the spreadsheets to see what impact this approach has on POT and it does increase (to 31%) but the number of bets is substantially reduced so could just be a quirck.

Welcome some feedback

Many thanks in advance

[Bhagwan]

200 races is not a bad sample.
If one is hitting 21% POT betting 4 horses, I would not change anything just yet.

Well done.

Be prepared for the run of outs that will come up, no matter how good the ratings.

What I do is stop after 9 outs in a row & wait for one of my selections to get up before re-commencing.
It works for me.

If one wishes to establish a clearer picture of how ones system may perform , one idea that works well is have 3 sample batches of 150 races.
If one batch shows a profit & the other 2 batches manage to break nearly square , the plan has potential.



The only thing that could kill the plan is short prices that are not in ones top 4 selections .

So its an idea to avoid any race where the second Fav in the race is not paying $4.00+

That way, one is only competeing with one thats poor value & a chance of getting up , not 2 horses of poor value with a chance of getting up.

Cheers.

[michaelg]

Sparky12, I would presume you are selective in the type of race for your strike rate of 68% and POT of 21%? and you are betting level stakes?

One of my friends had a friend who worked for Mark Read. He said that M.R. would total his assessed odds/percentage of his top three selections, and if he could get those odds or better he would then have a bet. For example, if the assessed total percentage of his top three selections came to 65%, and if another bookie/TAB were offering odds on those three horses that totalled 65% or better, he would then back all three horses. However my friend does not know if M.R. would juggle the amount bet on each horse or whether he would use those as suggested by his ratings.

I agree with Bhagwan - I would not be too eager to change anything yet.

[Sparky12]

Michael G - Yes, the system is restricted, in my case I focus only on Class 6 and above and I ignore 2yo and 3yo only events so exposed form is key, I'm using level stakes for my calcs ie: same amount on each of the four horses, race after race, irrespective of results - not actually betting anything as yet! V interested in comments re Mark Read's approach - the difference is he is assessing the odds of his top 3 picks based on his evaluation of horses in a specific race I suspect - in my case I'm simply assuming my 68% strike rate is valid and continues into the future and using that 68% "chance" of success as the basis for the value comparison - not sure yet that this is correct or valid hence the post.

Bagwan - thanks for the words of encouragement and advice - I'm on a losing role now following a fantastic run of results so great to be brought down to earth so quickly!

[crash]

Sparky,

21% pot return is a very good figure [exceptional] for backing 4 horses per race based on ratings. Boy, who's are those?

What is your average winning odds to reach that sort of figure? It would have to be quite high as even at $8 units win average your true odds for 4 runners is only evens. $4w would be money back and no profit. 68% SR is a hint I suppose but I can't figure it at the moment [I've got the flu]. Maybe someone here can[?]

[Sparky12]

Crash

You've got me worried now mate! Quick calc suggests average Div of $7.8, total races in this sample is actually 181. Not all races require all 4 horses to be backed - the system has a cut off price of $40 for a horse so anything over is a no bet (so sometimes only 2 horses backed and sometimes 3) - my ratings have thrown up a surprising number of ridiculous longshots which I don't bet. Having said that doing a box trifecta of my top 4 has produced a 7000-1 Trifecta of late and yes, I didn't invest! Infact a box trifecta on each race returns a POT significantly into 3 figures but if you exclude the 7000-1 tri then things looks very good but not the stellar retuns produced by including it! Quinnella results are good too interestingly enough.

If I was to restrict my bets to (say) only horses priced $16 and under, this would perhaps be more realistic test as longshot winners (and there's been a few) are eliminated from the results thus meaning less skewed result - the POT drops to around 9% (last time I checked) so maybe that is a good reality check.

The ratings are my own and fairly simply contructed so no rocket science here I'm afraid - hence keen to get feedback such as yours as I am a pessimist by nature and, like you, can't believe a 20%+POT would be achievable in a bigger sample by backing 4 horses. I'd love however to have the time to filter the four horses down to maybe one or two only with some proper form study as I do think that's got to be a way to some solid and regular profit but............

Cheers

[crash]

Well I'd keep doing what your doing for the moment but be very wary of your real odds. You should make sure your going to get at least evens on your shortest priced runner. A 3 horse bet would mean all of them are paying at least $6 min. If it's less, you are accepting odds-on. Not good in my book, but others are OK with it.

Swings and slides. Increase you chances, reduce your odds etc. I used to back 2 horses in a race for a long time but now I [mostly] just back 1 horse per race.

[Bhagwan]

Hi Sparky,
Try this with your past figures.
With the 1st & 2nd rated only bet them if paying $3.50+
With the 3rd & 4th rated only bet them if $8.50+

This should increase the POT.

2nd fav in race must be $4.00+ otherwise no-bet race.

Maybe make $28 as the cut off.

Cheers.

[Sparky12]

Thanks Bhagwan

This'll test my Excel skills for sure! Appreciate you giving this some thought - will run the numbers this weekend and see how I go and report back.

I reworked my ratings formulas to accommodate lower class races this week (Cl1-6 at country/provincials) - got 5/7 quinellas and 3/7 trifectas at Rockhampton yesterday based on boxing my top 4 rated - 6/7 winners as well - joy lasted about 1 minute when I saw the returns.......still, I enjoyed that minute immensly, considered giving up the day job, island in the sun etc etc....all good fun.

Cheers

[Dale]

Sparky i'm currently following my own ratings and they are producing figures that im very happy with,the strike rate of the highest rater isnt as high i would like but like you the top 3 or 4 is producing the goods,my suggestion to you would be to restrict yourself to races where the chances of a good divedend are higher.

Try 10 starters or more on a saturday and either 11 or 12 any other day,it works for me.

[Sparky12]

Thanks Dale. Good point re the number of starters, I've found that the results on the lower class tracks are good in terms of strike rate but the returns aren't enough to make it fly at the moment and mostly in small fields.


Good luck with your own system.

[Sparky12]

Dale - I meant to ask, what IS your strike rate for top rated and your top 4? I need to check latest data for my own ratings but my top rated doesn't deliver as much as I would have liked either...

[Dale]

Dale - I meant to ask, what IS your strike rate for top rated and your top 4? I need to check latest data for my own ratings but my top rated doesn't deliver as much as I would have liked either...

Hey Sparky sorry about the delay.

Ill have the answer for you in a few days,so far ive just been collecting results,dividends etc and havent got to that,ive got enough data for it to mean something now so ill start on it.

Off hand the top rater is hovering around the 20% mark,the top 3 around the 50% mark and the top 4 nudging 60%.

Cheers.

[Sparky12]

Dale - be interesting to see what the actual strike rates are, fyi my latest are as follows so maybe not too dissimilar to yours?

<TABLE style="WIDTH: 192pt; BORDER-COLLAPSE: collapse" cellSpacing=0 cellPadding=0 width=256 border=0 x:str><COLGROUP><COL style="WIDTH: 48pt" span=4 width=64><TBODY><TR style="HEIGHT: 12.75pt" height=17><TD class=xl26 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; HEIGHT: 12.75pt; BACKGROUND-COLOR: #ccffff" align=right width=64 height=17 x:num="0.1736842105263158">17%</TD><TD class=xl26 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right width=64 x:num="0.14210526315789473">14%</TD><TD class=xl26 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right width=64 x:num="0.21052631578947367">21%</TD><TD class=xl26 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right width=64 x:num="0.11578947368421053">12%</TD></TR><TR style="HEIGHT: 12.75pt" height=17><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; HEIGHT: 12.75pt; BACKGROUND-COLOR: #ccffff" align=right height=17 x:num>1</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right x:num>2</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right x:num>3</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: #ccffff" align=right x:num>4</TD></TR></TBODY></TABLE>

I'm running the same ratings formulas with some minor tweaks for low class races and the results for top rated are much better but overall returns (backing all 4) not so good even though the strike rate is better at 75%. Sample size (50 races) means this needs longer to settle down.

<TABLE style="WIDTH: 192pt; BORDER-COLLAPSE: collapse" cellSpacing=0 cellPadding=0 width=256 border=0 x:str><COLGROUP><COL style="WIDTH: 48pt" span=4 width=64><TBODY><TR style="HEIGHT: 12.75pt" height=17><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; HEIGHT: 12.75pt; BACKGROUND-COLOR: transparent" align=right width=64 height=17 x:num="0.37735849056603776">37.7%</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right width=64 x:num="5.6603773584905662E-2">5.7%</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right width=64 x:num="0.11320754716981132">11.3%</TD><TD class=xl25 style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; WIDTH: 48pt; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right width=64 x:num="0.20754716981132076">20.8%</TD></TR><TR style="HEIGHT: 12.75pt" height=17><TD style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; HEIGHT: 12.75pt; BACKGROUND-COLOR: transparent" align=right height=17 x:num>1</TD><TD style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right x:num>2</TD><TD style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right x:num>3</TD><TD style="BORDER-RIGHT: #ece9d8; BORDER-TOP: #ece9d8; BORDER-LEFT: #ece9d8; BORDER-BOTTOM: #ece9d8; BACKGROUND-COLOR: transparent" align=right x:num>4</TD></TR></TBODY></TABLE>
Bhagwan. Thanks again for your suggestions in your recent post. I reran the numbers based on your suggestions and sadly the POT reduced a little but I didn't manage to utilise the $4 second fave cut-off as you suggested (too hard for my poor excel skills) so maybe that explains it.

I did, however, run the numbers based on limiting bets to only backing all four when combined odds for top 4 rated were better than my current strike rate (64%) - the POT rose to 43% for 100 races. This seems to "prove" the theory I was keen to test in my initial post ie: the concept of value may apply to combined odds just as much as to individual odds but I still struggle with this a bit and suspect I may just have a lucky sample!

Anyway, v.grateful for your thoughts on this - will keep plugging away and see how it goes.

Cheers

[Bhagwan]

Thanks for sharing your observations.

I was reading some where that was one way of doing iy .
That is Betting all the selections Dutch betting or level stakes when the total available percentage is greater than our own because long term one should get in front.

The only thing that upsets the apple cart is when there are heaps of short payers comming in & that is why where the $4.00 2nd Fav rule , tries to address that issiue.

When ever one sees a $3.90 & less 2nd fav.
the race stats show that it will be won by one of the 2 Favs approx 45% of the time & thats not what one wishes for doing 4 horses.

Now that we are into a new season one will see a lot of this .
So my surgestion is save your gun powder for when the conditions are in your favour , rightly or wrongly.

Cheers.

[Dale]

Sparky

I think those figures of yours will settle down with more races,i mean that the top rater should start winning more than the 2nd,the 2nd more than the 3rd and so on.

If i was you i'd hold back on the tweaking untill you have a few hundred races,the changes you make will be more accurate that way.


I still havent finished the claculations for my top 4 but i do have some figures for my top rater

321 races(any race anywhere,hurdles Darwin you name it)
69 wins = 20.23%
returning $311.90 uni tab

There are some days i havent got to yet that i know will lift the return dramticly and i think eventualy the strike rate will end up around the 22% mark.

As i said earlier ignoring races with small fields lifts the results into good profit,when i finish the calculations ill post some stats.

Good luck.