Your AMMP is 3.227 which means at £100 stakes you would make £322.70 per month profit

Using the RANKING order and value in odds......I have STRATEGIES for :-

1. TOP 3 in market ( you can go to top 4/5 where JT/CFS etc)....

2. TOP 3 rated.

3. Laying those in lower part of ranking oder, where odds are value ( I will ignore certain rank positions re laying below the line ( line drawn at 50% + 1 of field size).

4. BACKING horses above the line that are within certain odds range ( you can get long runs of losers with this...but still profitable...theoreticaly...)



You can get days where a huge amount of points can be made.....and now and then days where you will lose a big chunk of points ( you can get a few bad days in a row....depending on which strategy you are using....)

Main thing is STAKE to take account of the "bad run" possibility....as they will happen due to there being SUBJECTIVE input in ratings....and I will have "off days" where brain not working well. Such days are few in number ...but if I have 2/3 days in a row where the subjective input is "wrong". the HITS can be very costly.

Generally though using RANKING ORDER and odds to determine bets produces PROFIT on every strategy mentioned above.

Sounds as if you are doing really well on your ratings BFF....wd


I'm still updating system re trainer stats.....BEFORE I can use system properly. It is such a boring "job" that I admit I have been more interested in keeping myself up to date re how race results are currently going.

Due to my personal circumstances, I just don't have time to use system properly...... however, I am finding that with small bets and ad-libbing from race to race can produce good profits, without resorting to HIGH STAKING.

I know that I should be using the system and strategies devised.....but it really takes a lot of energy and time to use system properly.....and currently i might continue just betting the way i am ( only started re HORSES just before CHELTENHAM using small stakes....and have made a profit from HORSE RACE bets amounting to £ 1,333.02 ( 12 days) .

Just doing what I am doing....I could up stakes by 5, or more,(could easily be financed)as there is sufficient liquidity for the type of bets placed.

It is frustrating not being in a position to use system properly.....but it is far easier, and less stressful, doing what I am currently doing. Finding it relatively "easy" to make £100 a day with small stakes......so say upping stakes by 5 on some races where I feel my edges are stronger could be done ....so will be doing so next week.

GL BFF with system and application of it ...though LUCK is irrelevant generally as you will know.

Cheers the nooooooooooooooooo

Great content in this thread and a mine of information to help all ventures accross the minefield of betting

bf - Where did you get your 322 a month profit from ?

£322.70 = £100L/s @ 3.227 APPM

bf_fananatic

Re. your Cheltenham ratings and mine which are shown below - several similarities so maybe we're using the same base source.

Bets: 81
Average BSP: 17.83
Wins: 13
Returns: 85.24 (BSP, 5% comm.)

Date	Time	Horse	Rank	Result	BSP	Returns
13/03/12	13:30	Steps To Freedom (IRE)	1	14th	8.91	-1.00
13/03/12	13:30	Cinders And Ashes	2	1st	12.00	11.45
13/03/12	13:30	Trifolium (FR)	3	3rd	8.53	-1.00
13/03/12	14:05	Al Ferof (FR)	1	4th	4.97	-1.00
13/03/12	14:05	Sprinter Sacre (FR)	2	1st	1.82	1.78
13/03/12	14:05	Cue Card	3	2nd	7.97	-1.00
13/03/12	14:40	Quantitativeeasing (IRE)	1	7th	7.80	-1.00
13/03/12	14:40	Our Mick	2	3rd	14.02	-1.00
13/03/12	14:40	Hold On Julio (IRE)	3	10th	7.25	-1.00
13/03/12	15:20	Hurricane Fly (IRE)	1	3rd	1.79	-1.00
13/03/12	15:20	Zarkandar (IRE)	2	5th	12.00	-1.00
13/03/12	15:20	Binocular (FR)	3	4th	7.59	-1.00
13/03/12	16:00	Scotsirish (IRE)	1	PU	5.03	-1.00
13/03/12	16:00	Dancing Tornado (IRE)	2	UR	14.29	-1.00
13/03/12	16:00	Garde Champetre (FR)	3	PU	19.07	-1.00
13/03/12	16:40	Quevega (FR)	1	1st	1.61	1.58
13/03/12	16:40	Violin Davis (FR)	2	15th	50.00	-1.00
13/03/12	16:40	Alasi	3	4th	51.74	-1.00
13/03/12	17:15	Triolo Dalene (FR)	1	11th	9.70	-1.00
13/03/12	17:15	Hunt Ball (IRE)	2	1st	9.00	8.60
13/03/12	17:15	Bless The Wings (IRE)	3	10th	11.81	-1.00
14/03/12	13:30	Four Commanders (IRE)	1	3rd	14.33	-1.00
14/03/12	13:30	Harry The Viking	2	2nd	9.40	-1.00
14/03/12	13:30	Alfie Spinner (IRE)	3	5th	10.50	-1.00
14/03/12	14:05	Sous Les Cieux (FR)	1	5th	7.00	-1.00
14/03/12	14:05	Monksland (IRE)	2	3rd	6.87	-1.00
14/03/12	14:05	Simonsig	3	1st	3.31	3.19
14/03/12	14:40	Grands Crus (FR)	1	4th	2.42	-1.00
14/03/12	14:40	Bobs Worth (IRE)	2	1st	5.70	5.47
14/03/12	14:40	First Lieutenant (IRE)	3	2nd	6.20	-1.00
14/03/12	15:20	Sizing Europe (IRE)	1	2nd	1.96	-1.00
14/03/12	15:20	Finians Rainbow (IRE)	2	1st	5.90	5.66
14/03/12	15:20	Big Zeb (IRE)	3	3rd	8.80	-1.00
14/03/12	16:00	Third Intention (IRE)	1	8th	27.00	-1.00
14/03/12	16:00	Balgarry (FR)	2	7th	7.45	-1.00
14/03/12	16:00	Get Me Out Of Here (IRE)	3	2nd	8.74	-1.00
14/03/12	16:40	Vendor (FR)	1	3rd	4.97	-1.00
14/03/12	16:40	Kapga De Cerisy (FR)	2	11th	36.43	-1.00
14/03/12	16:40	Une Artiste (FR)	3	1st	55.32	52.60
14/03/12	17:15	The New One (IRE)	1	6th	19.48	-1.00
14/03/12	17:15	Horatio Hornblower (IRE)	2	15th	14.16	-1.00
14/03/12	17:15	Village Vic (IRE)	3	13th	22.03	-1.00
15/03/12	13:30	Sir Des Champs (FR)	1	1st	4.70	4.52
15/03/12	13:30	For Non Stop (IRE)	2	3rd	10.50	-1.00
15/03/12	13:30	Cristal Bonus (FR)	3	PU	8.83	-1.00
15/03/12	14:05	Restless Harry	1	11th	9.28	-1.00
15/03/12	14:05	Rick (FR)	2	6th	44.00	-1.00
15/03/12	14:05	Reindeer Dippin	3	5th	82.14	-1.00
15/03/12	14:40	Rubi Light (FR)	1	5th	7.43	-1.00
15/03/12	14:40	Somersby (IRE)	2	7th	5.81	-1.00
15/03/12	14:40	Riverside Theatre	3	1st	4.39	4.22
15/03/12	15:20	Big Bucks (FR)	1	1st	1.87	1.83
15/03/12	15:20	Mourad (IRE)	2	6th	50.72	-1.00
15/03/12	15:20	Voler La Vedette (IRE)	3	2nd	35.79	-1.00
15/03/12	16:00	Notus De La Tour (FR)	1	F	9.04	-1.00
15/03/12	16:00	Hectors Choice (FR)	2	17th	16.50	-1.00
15/03/12	16:00	Giorgio Quercus (FR)	3	14th	23.40	-1.00
15/03/12	16:40	Start Me Up (IRE)	1	5th	13.50	-1.00
15/03/12	16:40	Adams Island (IRE)	2	19th	2.40	-1.00
15/03/12	16:40	Up The Beat	3	4th	10.19	-1.00
16/03/12	13:30	Grumeti	1	3rd	6.37	-1.00
16/03/12	13:30	Baby Mix (FR)	2	9th	8.40	-1.00
16/03/12	13:30	Countrywide Flame	3	1st	46.00	43.75
16/03/12	14:05	Raya Star (IRE)	1	13th	10.50	-1.00
16/03/12	14:05	Citizenship	2	14th	7.40	-1.00
16/03/12	14:05	Sailors Warn (IRE)	3	3rd	22.37	-1.00
16/03/12	14:40	Boston Bob (IRE)	1	2nd	2.50	-1.00
16/03/12	14:40	Rocky Creek (IRE)	2	8th	18.10	-1.00
16/03/12	14:40	Brindisi Breeze (IRE)	3	1st	9.00	8.60
16/03/12	15:20	Long Run (FR)	1	3rd	2.86	-1.00
16/03/12	15:20	Kauto Star (FR)	2	PU	4.50	-1.00
16/03/12	15:20	Midnight Chase	3	7th	17.46	-1.00
16/03/12	16:00	Enter Paradise (IRE)	1	UR	44.00	-1.00
16/03/12	16:00	Bradley	2	8th	38.00	-1.00
16/03/12	16:00	Keenans Future (IRE)	3	12th	216.54	-1.00
16/03/12	16:40	Molotof (FR)	1	13th	17.18	-1.00
16/03/12	16:40	Make A Track (IRE)	2	4th	19.97	-1.00
16/03/12	16:40	Imsingingtheblues (IRE)	3	18th	27.69	-1.00
16/03/12	17:15	French Opera	1	F	26.00	-1.00
16/03/12	17:15	Kid Cassidy (IRE)	2	12th	6.26	-1.00
16/03/12	17:15	Astracad (FR)	3	7th	7.80	-1.00

N.B.  Removing the two longest priced winners (which amounts to just 15.4% of the total number of winners) alters the picture completely and would have resulted in a level stake loss. 

.BFF got the winners in their selections -----should be ---.BFF got the winners in the selections

Glasgow Brian - any system which uses market position of the horse is going to be a nightmare when it comes to putting your money down in reality.

You are looking for the fav in your system. But what horse is fav?  OK a lot of the times it is obvious and doesn't change, but what if there are two or three at the top of the market, flip-flopping about. What if one horse if fav on BF but not on the show? Do you lay it or not?  Also when do you lay?  10am your "fav" could be out with the washing as it drifts come race time.  Just before the race - then you have to stick at your PC right until the off.

It's all very well looking back in time and saying the fav did this, that or the other - but that is when you know what the fav was.  Even worse if you have a system that says look at the top 3 horses in the market as somebody else said later on this thread.

The forecasted BSP at time form is pretty accurate but a pain in the rear to convert in its current format and although you cant always be sure what is going to be favorite you can get a good idea of the price band that all the runners will occupy, non runners affect the market much more but you can use auto-betting apps to have a reserve set up using a ranked set-up
The bet engine allows this type of bet automation and I can recommend it totally having tested it out.

The only systems apart from trading that truly work or those that have underlined
types of form or data that increase the strike rate above or below the implied chance via better than standard odds compilation and other backers changes to the market when live.

But what the real question boils down to is that are the methods that are used to locate
mis-priced selections in reality genuine valued calculations or products of the continually shifting sands of natural variation that move peaks and troughs of returns across the spectrum
of available data that can mislead.

To find this answer and to answer your question Brian , 1 to infinity may be either profitable or a loss as natural variation smooths out.

LOL FAFH try this one for size....All theories and calculations used to work on the subject matter of probability of an event are only rational and thus useful at certain points in time and beyond certain points in time they become void.

E.G. all markets in betting of sporting events that involve chance are meaningless once the race is finished at that point in time, as the winner has not even 100% chance of winning though it has won as it cant be a predicted chance anymore but a conclusion of the probability calculation that itself is now redundant in its useful form.

The logical conclusion of the above observation is that

1 X probability = 0 x result and also 0 x probability = 1 X result

But how can this be as 1 cannot equal zero unless the 2 separate multiplying factors are them
selves opposite thus 1 X 0 can equal 0 x 1, 1 and zero are in fact opposite as one exists and zero does not.

So basically you cannot have probability and result in the same time as both only exist in separate times, one the future and one the past which are opposites.

Well if you have read up to this point and wondered what the hell is this guy on about well
its a lesson in pure observation and logic and also leading up to a powerful quotation which I want all you Betfairians to remember.

If you want time to spend on a result then spend time on probability

And for sanity's sake don't waste time understanding someone else's version of probability
like you have just done now and when you look at the average tips and prices that are floating around.

So after all this, what is the minimum sample size needed to know whether you are a winner of not please?

There isn't a simple answer to that - in theory you can fluke an unlimited number of results plus it depends on the way in which your samples are derived.

That said; if you're looking for a rule of thumb - at odds of 3-ish you'd probably want a couple of thousand. At significantly longer or shorter odds you'd need more. Whatever you think you need - it's always more.

There are various statistical tests you can do to assess the strength of a correlation but personally I've never found them that useful, there are however a couple of simple heuristic tests that I use all the time:

1: Look at the neighbors - IE those that almost met your criteria but didn't quite. Say you're scoring selections out of 10 and backing those that scored 8 or above, how did the 7s do? If your strategy is good you should see a sloping improvement across the score range from negative to break even to positive. The 10s should do better than the 9s and so on. If you're seeing peaks and troughs you're either not onto anything or you don't have enough data. How you construct the neighbors depends what you're analysing. Say you've got an indicator that's good for 5f races (and the distance is somehow relevant) it should still roughly hold up for 6f, a bit less for 7f and so on.

2: Divide your data. The stronger a trend is the smaller the data set needs to be in order for it to be evident. If you divide your data in half you should see the broadly the same trend in both sub-sets. The better your indicator the more times you can make the division whilst still observing the trend in each sub-set.

3: Check the opposites. If your indicator highlights good value to back then selections with the opposite characteristics should be value to lay. If they're not then you've got to ask yourself why.

4: Check your reasoning and your process. First; just ask yourself honestly whether what you've 'found' is probably just b0llocks. If you cant explain convincingly WHY a trend exists then you should treat it with great skepticism. Ditto if what you've found is incidental - IE you were expecting or looking for something else. If you came up with a theory, made a prediction then observed it then that's great. If you just went fishing or back-fitting then you not only need a bigger data set but a different data set - i.e. one from the future not the past.

Above all - decide your methodology BEFORE you do your analysis. Otherwise all you'll do is keep going through different techniques until you find one that makes it look like you're going to be rich. That won't work. Not ever.

Thanks for your replies.

Beginning of 2015 I decided to take things seriously with the horses, after 30 years of failing & subscribed to a ratings/tissue service and started keeping results. At one point I was £12k up, then lost the lot. I finished the year £3.2k up. Last year exactly the same thing happened I was £6k and again lost the lot. It forced me to look back at my records and it seemed to me the races that were killing me were big field competitive races, handicaps, (the type of race Pricewise tips in) backing more than one in a race and backing ew. Beginning of August last year I decided to change my strategy.

Now, I only consider races where the horse at the head of the betting on the tissue is 4-1 or less, so open races where they are 9-2+ the field on the tissue I use are an instant swerve for me now, and I only back horses that are 4-1 or less on the tissue to try and cut down the horrendous losing streaks I have had (and still have) I very rarely back more than one in a race and I never back ew.
Obviously although the tissue price on a horse I'm looking at is 4-1 or less I'm looking to equal or beat these odds when I place a bet. With a nearly a year behind me here are the results:
Month: in (returns £) out (amount staked £)brackets: profit/loss Return On Investment strike rate
Aug 2016 in: 10618  out: 9470 (+1148) ROI: 12% 11/68 16%
Sept 2016 in: 15257  out: 13975 (+1282) ROI: 9% 21/102 21% in: 25875  out: 23445 (+2430) ROI: 10% 32/170 19%
Oct 2016 in: 13364 out: 10490 (+2874) ROI: 27% 24/100 24% in: 39239  out: 33935 (+5304) ROI: 16% 56/270 21%
Nov 2016 in: 8825  out: 5850 (+2975) ROI: 51% 15/57 26% in: 48064  out: 39785 (+8279) ROI 21% 71/327 22%
Dec 2016 in: 14471 out: 15200 (-729) ROI: -5%  24/127 19% in: 62535  out: 54985 (+7550) ROI: 14% 95/454 21%
Jan 2017 in: 28520  out: 21976 (+6544) ROI: 30% 44/168 26% in: 91055  out: 76961 (+14094) ROI: 18% 139/622 22%
Feb 2017 in: 21716  out: 22721 (-1005) ROI: -4% 36/159 23% in: 112771  out: 99682 (+13089) ROI: 13% 175/781 22%
March 2017 in: 60901  out: 54198 (+6703) ROI: 12% 60/243 25% in: 173672  out: 153880 (+19792) ROI: 13% 235/1024 23%
April 2017 in: 47947  out: 55752 (-7806) ROI - 14% 34/191 16% in: 221619  out: 209640 (+11979) ROI: 6% 269/1215 22%
May 2017 in: 71209  out: 70264 (+945) ROI : 1% 52/237 22% in: 292828  out: 279904 (+12924) ROI: 5% 321/1452 22%
June 2017 in: 38766  out: 37978 (+788) ROI: 2% 25/131 19% in: 331594  out: 317882 (+13712) ROI: 4% 346/1583 22%

The figures in italics are rolling totals. At the end of June I had staked £317882 which returned £331594 for a p/l of £13712 - a ROI of 4%. I've backed 1583 horses, 346 of them have won for a 22% strike rate. What I'm worried about is, my strike rate has remained pretty consistent around 22% but my ROI is steadily decreasing. Am I reverting to the mean, have I just been lucky, will I eventually become a loser or is 4 maybe 5% a realistic roi and I should continue. Any thoughts would be appreciated, and feel free to ask any questions. TIA

Roughly 2 or 3 hours a day, including putting bets on. Normally up between 7 and 8, turn on the computer and get started. Used to be 4 or 5 hours, but by avoiding races where they are 9-2 plus the field I've cut down the number of races I need to study, which is good because I ended up burnt out. No, it's not full time, I have a business to run, but it's seasonal and I have a lot of time on my hands especially during the week and for 4 months in the winter the business closes down.

I must read this thread in detail as it looks like some people know stuff.
For a long time I've been working on pedigrees.
At the start of 2017 I wrote programs to analyse pedigrees.
My sample size is 159,000.
It was a bit of a moment to look at the results for the first time to see if years of work were pointless or not.
Results were favourable.
It isn't for betting, although I might use it as a guide later.

Imagine you are looking back at past races and discover a profitable betting strategy, with a reasonable basis for why it makes sense, and so on. You might have found an edge, fantastic! The problem is, edges tend to disappear over time, as other punters grow wiser. The flip side of this is that the further back in time you look, the larger the edge is likely to be. So if you test your strategy over more races, say the last two years of racing, you could find it produces better results than it does over just the last year. You might think that this is great news and confirms your strategy, and it does, but it can also make your edge look far bigger than it really is.

This is a fair point, though I would say it's an easily avoidable trap if you're graphing your data - which you DEFINITELY should be. In this case you'd see the trend emerging visibly over time if you binned (grouped) all your results by date and graphed the mean result.

More widely I find graphical approaches to data analysis tremendously useful. Any phenomenon that's strong enough to exploit will almost certainly be readily apparent on a graph. It's my "go to" approach for edge discovery. I only really use more advanced statistical tools to work out how indicators relate too each-other, all the foundational work finding out what's relevant and comprehending how the markets work is done with graphs. I have tried other approaches but I've never got them to yield much. I even hired a statistician to look into some data for me, he was a university professor who specialised in machine learning so he was qualified, but even he didn't come up with anything I didn't already know about.

Peter May
@PeterMayRacing
·
6h
My latest blog piece is already on the ATR website.  It covers an approach based 3yos in all age handicaps which I referred to in my last book back in 2013.  Good to see it still working.