The Ratings Thread

[dave jay]

If you email it to me I’ll have a look ..  or give a fuller explanation, it shouldnt be a problem.

[Have fun]

Mine for today<br>Madison du berlais  92 =22.96%<br>Flying enterprise  81 =18.68%<br>Supreme prince  80 = 17.51%<br>Nozic  79 = 17.90%<br>Marcel  79 =17.90%<br>Fundamentalist  78 =17.51%<br>Kalco mome  65 =12.45%<br>No full  62 = 11.28%<br>Kelrev  59 =10.12%<br>It takes time  55 = 8.56%<br>Ichi beau  34 =0.39%

hope we all do better this time<br>

(Edited by Have fun at 12:46 pm on Mar. 3, 2007)

[Formath]

3.15 Newbury<br>Ratings<br>15 Marcel 4/1  (11/2 lost) 20pts (19.697%)<br>12 Flying Enterprise 11/2 (11/4 no bet) 15pts (15.152%)<br>9   Supreme Prince 17/2 (10/1 lost) 10pts (10.606%)<br>8   Nozic 10/1 (7/2 no bet) 9pts (9.091%)<br>8   Kalca Mome 10/1(12/1 lost) 9pts (9.091%)<br>7   Madison Du Berlais (7.576%)<br>7   Kelrev (7.576%)<br>7   Fundamental (7.576%)<br>7   It Takes Time (7.576%)<br>5   No Full (4.545)<br>3   Bold Bishop (1.515)

Result -39pts

Hopefully, I won’t be out of step today lol. I do normally use prize money won as my class filter (because it is universally available) my hcap class note was just to indicate the ‘official’ opinion. Take this race today:

Official hcap rating 153, hcap chase over 20f on going f/c as Soft. Considering those with the best recent form where you could expect the win to arise, this is the best win OR they have achieved:

144 Marcel down 11lbs, over 17G, feasible weight<br>138 Nozic up 6lbs, over 23Hvy, how good is he?<br>137 Flying Enterprise down 14lbs, over 17G, progressive hcap OK<br>137 Madison Du Berlais down 2lbs, 16S, weight at the D?<br>135 Kalca Mome up 1lb, over 17S, most wins at 16f?

(Edited by Formath at 4:38 pm on Mar. 3, 2007)

[dave jay]

This is our combined rating ..

Supreme Prince,10.17%,9.83,9/1<br>Nozic,11.50%,8.70,8/1<br>Marcel,14.25%,7.02,6/1<br>It Takes Time,6.55%,15.27,14/1<br>Flying Enterprise,16.49%,6.07,5/1<br>Kalca Mome,9.04%,11.06,10/1<br>Madison Du Berlais,9.82%,10.18,9/1<br>Bold Bishop,3.17%,31.51,33/1<br>Fundamentalist,5.76%,17.36,16/1<br>No Full,6.32%,15.83,16/1<br>Kelrev,7.55%,13.25,12/1

.. interesting!

[dave jay]

Well done have fun, your top rated won .. 12/1

Our combined ratings also picked it .. <br>Supreme Prince,10.17%,9.83,9/1 No Bet<br>Nozic,11.50%,8.70,8/1 No Bet<br>Marcel,14.25%,7.02,6/1 No Bet<br>It Takes Time,6.55%,15.27,14/1<br>Flying Enterprise,16.49%,6.07,5/1 No Bet<br>Kalca Mome,9.04%,11.06,10/1 -11<br>Madison Du Berlais,9.82%,10.18,9/1Won 12/1 +120<br>Bold Bishop,3.17%,31.51,33/1<br>Fundamentalist,5.76%,17.36,16/1<br>No Full,6.32%,15.83,16/1<br>Kelrev,7.55%,13.25,12/1

+110 pts<br>Well done all !!<br>:biggrin:

[cormack15]

Well done lads – great thread

[carlisle]

Hi dave

I have found the Excel function called RANK.  So it should be easy enough to write a vdw spreadsheet now.

byefrom<br>carlisle<br>

[dave jay]

Carlisle, the RANK function will assign a numerical value within a list which is what I was thinking you wanted. It helps to set the formula up like this ..

=RANK(1,B$3:B$13)

.. if you pin the positions on the list when the size of the list changes you change the top cell and copy it over the others.

Alternatively and  preferred, but a bit harder. Modify the input data to default to a very small value for no runner. So if your list is set up for 20 runners and you only have 10, instead of getting the N/A error (for null value), which messes up the RANK function, it is always 0 and therefore bottom ranked. That way you dont have to dick around with the sheet everytime.

Good luck.

[Pompete]

Rated up the 4.50 Bangor just for fun if anyone interested:

Cantgeton  123  (18%)<br>Rare Coincidence  120  (14%) Non-Runner<br>Gardasee  120  (14%)<br>————————————–<br>Livingonaknife Edge  119  (13%)  1st 8/1<br>Paperpprophet  119  (12%)  Non-Runner<br>Caliban  118  (11%)<br>Our Jasper  117  (10%)<br>Speed Venture  114  (6%)  2nd 5/2<br>Contact Dancer  110  (1%)

Reflections on result:<br>On a positive note the good priced winner came from the top three rated therefore meeting my own personal criterion. Also the ’tissue’ indicated a vaule bet to be had.

However, the ratings failed to indicate or reflect not only the manner in which the victory was achieved – very comfortably, but also how the field would be placed with the top two rated beaten by the bottom two rated (Contact Dancer fell) althought excuses could be made for Speed Venture. However, Cantgeton (Top Rated) appeared to be given an easy ride by TS and was never put in a position to pressure the leaders and Gardasee was apparently (first run in 355 days) "not fully fit" – see main thread. Nevertheless, neither would have beaten the winner.

Overall Verdict: Relatively Pleased.

Note: I hope this indulgence in writing up my post race views is o.k. The hope is that at sometime in the future it may help in deciding whether any progress is being made. The problem as always is ‘hindsight is a wonderful thing’ and I shall have to take to avoid falling into this trap.

(Edited by Pompete at 6:54 pm on Mar. 4, 2007)<br>

(Edited by Pompete at 7:34 am on Mar. 5, 2007)

[Formath]

OK Pete I will join you.

4.50 Bangor<br>Ratings<br>15 Speed Venture 7/4 (5/2)  36pts bet lost<br>12 Our Jasper 5/2 (10/30) 28pts bet lost<br>8   Cantgeton 5/1 (8/1) 16pts bet lost

5   Gardasee<br>4   Livingonaknifedge<br>3   Caliban<br>0   Contact Dancer

Result -80pts<br>(just remembered why I seldon bet in hcaps and never in hcap hdl races lol)

(Edited by Formath at 6:31 pm on Mar. 4, 2007)

[Pompete]

Cheers Alan

Shame about the non-runners. Interesting you’ve got Speed Venture top rated as I don’t know what to make of his last run at Haydock which I rated on form as 7Ibs higher than 12 other formlines I have for him and a whooping 35 speed points higher. On those figures I’d place him as 5/2 fav. However, a line through Karathanena (runner up) and Wizard of Us (thrid) makes me question the G/a.

Such is racing. :(

[dave jay]

Good Luck today .. a bit too busy to get stuck in, normal service should resume tomorrow .. :cool:

[Pompete]

I shall be posting up a couple of hurdle races tomorrow (a handicap and a seller at Newcastle, Formath – theory is if I can do it there I can do it anywhere, New York, New York)

But anyother suggestions? Happy to have to go.

[Pompete]

2.30 Newcastle

Stormy Beech   107  3rd 17/2<br>Turbo   107<br>————————–<br>Planters Punch   105<br>Zeitgeist   105<br>Culcabock   105   2nd  4/1<br>Flake   103 1st 9/2<br>Wenlocks Wonder   102<br>Armentierers   101<br>Neptune Joly   95+ (Minimum Novice Form Available)

<br>4.00 Newcastle

Cha Cha Cha Dancer   102   1st 8/1<br>Over To Joe   98?<br>Lucky Arthur   97<br>———————————-<br>Oleolat   95<br>Piccolomini   93<br>Formal Cliche   91   2nd 17/2<br>Named At Dinner   90<br>Winds Supreme   87<br>Minister Abbi   87   3rd  7/2<br>Obara D’Avril   84

Relections<br>I was too busy to watch the races so unable to relefect on running, however Cha Cha Cha Dancer won very comfortably (listened on Betfair radio) which is pleasing as it is reflected by the ratings – but that’s about the only thing that is!

<br>

(Edited by Pompete at 4:34 pm on Mar. 6, 2007)

[dave jay]

2:30 Newcastle<br>CULCABOCK .. 7/2<br>STORMY BEECH .. 4/1<br>PLANTERS PUNCH .. 9/2<br>ZEITGEIST .. 6/1<br>WENLOCKS WONDER .. 8/1<br>ARMENTIERES .. 9/1

Good luck all .. stakes as percentage of price<br>

[Formath]

Pete,<br>IMO if you can do it anywhere with hcap hurdles you deserve to become revered as St Pete, patron saint of punters lol.

2.30 Newcastle<br>Ratings<br>12 Culcabock 10/3 (4/1 lost) 23pts lost<br>9 Wenlocks Wonder 5/1 (9/1 lost) 16pts lost<br>8 Stormy Beech 6/1 (13/2) 14pts lost<br>7 Armentieres 7/1 (25/1) 12pts lost<br>5 Flake<br>5 Zeitgeist<br>4 Turbo<br>2 Planters Punch

Result -65pts

Debrief<br>I am only doing this today because it is of interest (to me anyway lol) as I can’t abide back-fitting, cherry-picking and hindsight as a way of proving a betting approach, normally.<br>I am on the weight-side-is-critical side of the argument regarding handicap races, coupled together with winning performance. This race was  0-110 but went off as 108. Checking the OR for wins and weight-change this is the result:

Flake down 4lbs from a 111 hcap chase win = possible 115<br>Turbo up 13lbs from a 110 win = possible 97<br>Culcabock up 13lbs from a 100 win = possible 87<br>Stormy Beech down 14lbs from a 94 hcap chase win = possible 108<br>The remainder of the field was not worth bothering with. Not sure exactly what it proves except that Flake was on a good hcap mark and overall it was a good shortlist.

Dave,<br>Something for you to get your teeth into on this dismal racing Tuesday as it seems to be in your area of expertise. It is a precis of correspondence I had with a pro-backer several years ago and lends itself to my approach. I would be interested in comments.

MULTIPLE BETTING BANKS

It may come hard to dedicated form students, but the simple fact is multiple banking works regardless of the number of runners covered, form is meaningless.  If it was not for the logistical problems you could back every runner in every race and still profit. Very few listen let alone understand!

When employing retrieve staking the biggest mistake made is to attempt to retrieve all losses.   Look at the facts, if you are anywhere near competent you ought to be able to identify a selection method that if combined with level stakes would provide no worse than a 10% loss.   Just backing favourites forecast between Evens and 2/1 should provide that.   Consider that, long term for every 100 points the expected loss is about 10, so why is it necessary to retrieve every point lost at once – it isn’t!

There is the problem of the losing run, but with a reasonable size bank we know it is not all going to be lost often.   Months could go by without a losing run that will break the bank, although eventually one will arise. How is that to be dealt with?

Simple, spread the risk to cover more than one horse.  If the risk of a losing run is small with one bank. How much smaller is the risk of a losing run in two banks, or three, or four banks at the same time? If losses start to creep up share the losses with the other banks and that will reduce the stakes in any one bank, at the same time increasing chances of finding a winner.

For example, two banks were run on a racing campaign.  The average price for bank A 6/4 and B 5/2.  Bank A had a losing run of nine bets yet overall there is a small profit.  If one bank had been used with a normal retrieve staking plan there would have been a loss, but using two banks to spread the risk makes the campaign profitable.

In the European football championship three banks were used, one for each outcome – home win, draw, away win and a fourth bank, referred to as the profit bank.   Of the three working banks two of them are always in deficit and they show a nett loss between them.  The profit comes from the profit bank.

Each match one of your banks has to achieve the aim owing to the fact that one of the three results has to transpire.   The aim for the campaign was £20 from a £500 bank.   Each match one bank achieved the aim, and about 80% of the time a little of the previous losses.  £20 won goes into the profit bank and any remainder comes off the previous losses of the bank that had the winner.

After 25 matches you are guaranteed a profit because 25 x 20 = £500.   From that point on any returns added to the profit bank are guaranteed nett profit.   The idea is to keep the three working banks balanced by spreading the risk.   Although they are always negative they still have enough in them to provide stakes to produce profit in the profits bank.

It is trickier with horse racing as you do not have a guaranteed winner each time.   One way round this is to concentrate on small fields (one bank per outcome); another is to look in areas where the odds compensate.

If you backed the first four in the betting in handicap races of  ten runners or less, your expected strike-rate would be about 70% with better odds of about 3/1 on the favourite.   This takes the pressure off as the four banks you are running are able to absorb the losses better.

Another way may be to back all the selections on VDWs narrowed field via the mechanical process advised.   Yet again one bank per outcome where the odds again are likely to be better as the races are generally more competitive owing to the class of races concentrated on.

(Edited by Formath at 5:53 pm on Mar. 6, 2007)

[dave jay]

Interesting post Alan .. I’m not going to reply to this in a straight forward way because it’s a complex subject that requires some proper analysis.

As you and I know, there are three things that separate successful punters from unsuccessful ones, they are; being profitable overall with your selections, backing your selections at the right price and staking correctly. None of these three things can be separated out from one another in reality as they form the whole picture.

Firstly, if you take the case of being profitable with your selections and backing at the right price .. unless you are e/w thieving or backing horses early on, on Betfair and then trading them out for a profit before the off, you have to have an edge. Having an edge means that your horses win in accordance with their true chances of winning but the price that you get about them is better than their true chances, if that makes any sense.

Onto retrieval staking, this does not work in statistical models. If overall, you make a loss of 10% then no matter how you stake you will end up with -10%. The effect of retrieval staking on your bank is to create a roller coaster effect, with wider fluctuations in the bank size, but the end result is the same. In addition to this it can lead to gambler’s anon. due to the added pressure and psychological problems it can create. What you are doing with any form of retrieval staking is gambling on the sequence of events, or the order your winners will appear in. You could be lucky and make a profit for a year or so, but the statistics will eventually catch up with you. The fellow that wrote that piece has failed to mention that his total values across all of his banks, at any given time, will be the equivalent of what his returns would be if he had simply backed every outcome to level stakes, be that money he is chasing or money that is owed.

Any staking plan that is not a stochastic process of some sort is gambling and best avoided.

What a stochastic process does, for staking, is take into account the overall size of your bank and calculates the next bet in accordance with that. Many punters get around this by creating plateaus to bet from, for example reduce stakes if the bank decreases by 20% or raise the level of the stakes if the bank increases by 20%. The correct way to stake is to adjust the stake sizes after every bet, but sometimes this isn’t practical. Staking this way locks both profits and losses out, whereas retrieval staking locks losses in.

Kelly staking is a stochastic process. It works something like this .. you work out the price you think the horse should be and compare this with the price you can get. Suppose your tissue has a horse at 2/1 (33%) and you can back it at 4/1 (20%). You would stake (33% – 20%) 13% of your bank on this selection. This is full Kelly and things can get a bit out of hand, a turned down version of this and the one I prefer, is this .. (33% – 20%) 13% divided by the odds on offer. So, the size of bet I would be staking here would be 13% / 4 = 3.25%, to WIN 13%, of my total bank value.

I’ll stop there I think !!<br>:cool:

[Author]

Posts