Betfair Free £25 bet. | Ten golden rules of betting

Ten golden rules of betting

Long-time successful punter Simon Rowlands offers up ten of the finest nuggets of gambling wisdom. Please note these are not THE ten, nor are they confined to horseracing. Read on...

Recent discussions on this site got me thinking about what is good practice and what is bad practice when betting. It is best not to be especially prescriptive in such matters - to a degree, if it works for you then it works - and the following is intended primarily as a starting point for discussion. Post your own thoughts in the "Comments" box below.

1) Value - the number one rule. Back enough selections at odds bigger than are merited by their chances and you have cracked it. Most other rules are ways of bringing this to pass or of ensuring that you capitalise on this advantage. Value can involve backing something at [1.1] when it should be at [1.01] or at [40.0] when it should be [20.0], or of laying the bets to others when the opposite applies. The problem in betting on horseracing and other sports is that it is not possible to know for sure what represents value in an individual instance, as would be the case with cards or coin tosses. Nonetheless, value as a concept is an inescapable truth. Betting is, in effect, an extended experiment in determining how good you are at judging value.

2) Take responsibility - if something goes wrong with your bet, do not automatically blame the jockey (who you knew would be riding your selection), the tipster you followed or the statistics you used (in both cases, it was your choice). Most of the best punters I have known have considered whether they have made a mistake with their bet - even when it has won - and sought to learn from the experience. They have taken responsibility for their actions, in other words. Bad punters almost always do not.

3) Fear and greed - thanks go to Andrew Black, founder of Betfair and one-time professional punter, for this gem of wisdom. As "Bert" once said, good punters have the right balance between fear and greed. Most bad punters are too greedy, trying to win too much at one go or increasing their stakes the moment they become successful (or greedily chasing their losses when they are not). Then again, I am too cautious by half. I frequently have not pressed what advantage I do have and am forever looking gift horses in the mouth. Try to find the balance appropriate to the type of betting you are engaged in.

4) Discipline - do not give in to the temptation to bet on more events than is suitable given your knowledge and preparation. Remember that you have the power not to bet at all. Do not chase your losses or get carried away with your wins. If in doubt, get up, walk away and come back another day. Similarly, do not cut corners in your preparation - betting tends to punish those that do - and discipline yourself to bet only when you are in a position to do it well.

5) Adapt or die - even in the eight or so years that Betfair has been in existence, things have changed markedly. More sophisticated forms of analysis are available and where the edges are have changed greatly. I started off laying books, primarily in the place market, moved on to the win market, did a bit of in-running betting before it became quite so popular, spent some time (when I was busier with other things) betting on individual race winning-distance markets in their infancy and have reverted to fairly old-fashioned on-the-nose backing (and laying) in recent times. Betfair provides the punter with huge choice and flexibility. It is my experience that, on Betfair at least, when one door closes another usually opens. Do not get stuck behind a closed door.

6) Play to your strengths - while point five urges the need to be versatile and adaptable, you should never bite off more than you can chew and should always play to your strengths. My knowledge of horseracing has been built up over decades, whereas my knowledge of Big Brother is short-term and superficial. It does not prevent me from betting on the latter, but I am well advised to know my limitations in this area. By a similar token, I feel comfortable with numerical analysis of horseracing - ratings, sectionals and the like - but far less so with paddock inspection, for instance. Paddock inspection is important, but I do not immediately leap on something that catches my eye before a race when others are better placed to do this.

7) Staking and money management - the Kelly Criterion is, to the best of my knowledge, the last word in correct staking (see previous article). Even if you do not act on every last symbol of the Kelly formula (I do not), be guided by the basic principles: you should stake a proportion of your bank in accordance with both the edge you perceive you have and the likelihood of the event happening. More generally, bad punters tend to vary their stakes alarmingly, ramping them up markedly the moment they get a few winners (or, in some cases, a few losers). Set aside money for betting and manage it so that you are in a position to capitalise on opportunities when they arise. Be consistent: play the long game.

8) Pace yourself - if you are betting at all seriously make sure you find time for other things, and do not become so obsessed with turning a buck that you burn yourself out. I used to give myself incentives and rewards for good behaviour and good results when I was betting for a living. I operated better if I had some "down time" a few times each week, and I made sure I took holidays. I imagine something similar applies to everyone.

9) Question received wisdoms - betting, and in particular horseracing, is full of received wisdoms. Some of them stand closer inspection (the primacy of "value" is a good example) but many do not. It is a human trait, encouraged by evolution, to look for patterns, and it is a human weakness, punished by experience, to see patterns where they do not exist. Pace biases, draw biases, track biases, golden highways, yardstick handicapping, trends, favourite jockeys, five-year-olds in the Champion Hurdle, going preferences of the offspring of sires: some good has been spoken and written about each of these but even more balderdash and piffle. Be sceptical, without closing your mind. There is a lot of mileage in going against the crowd.

10) There are few golden rules - betting on racing has had an enduring appeal in no small part because it is pitched at just the right level of complexity for the human mind. It has a habit of punishing the grimly deterministic and the fickle backer of hunches alike. Remain flexible: use judgement and intuition where it is appropriate.

24 September 2008 / About Simon Rowlands

Tags: Andrew Black, Betting Strategy, Big Brother, paddock inspection, ratings, sectionals, Value /

Comments (8)

Hi Simon,

One thing mystifies me about all this talk of 'value'. As you point out there is no objective way of determining the probability that a horse will win.

How do you come up with a price? Can you use any factors you like, i.e. speed ratings or trainer stats.

Cheers,

Stu

Stu | 24 September 2008

Good Article. Feel better for having read it. I would add two more golden rules to the list --never mix alcohol and betting --- jockeys and horses and trainers all tend to repeat themselves for the better or the worse.

paul jones | 24 September 2008

Hi, Stu.

I wrote a piece about pricing races up on here some time ago and more recently one about value. You can look around for them or the betting.betfair.com editor may be able to embed a link to those articles in the above one (?).

In terms of my own way of deciding what represents value I prefer to have in mind a reliable set of ratings but then to use my judgement to compile a fair-price tissue involving consideration of whatever factors seem relevant.

It would be difficult, but not impossible, to write the business rules required to automate this, as each race is different, as is each horse. It is something I certainly intend looking into more in due course.

I know of people who bet entirely round ratings and numbers derived from stats. In order for it to work, the methodology needs to be very sophisticated in my opinion.

My advice to anyone starting out in this area would be to try pricing up fairly easy races and to compare your conclusions with the odds on offer on Betfair. It is one area in which most people definitely get better with practice.

Thanks very much, paul.

I agree, mixing alcohol (or other substances that affect judgement and a sense of proportion) with betting is not a good idea unless all you are looking for is entertainment, at a price.

I agree also about the repeated behaviour, though with a few reservations. While it is not difficult to spot trends relating to a trainer who has been training for several years and sent out lots of runners as well as winners, it is not always possible to do so with a horse itself. They sometimes run only a handful of times in a career, and they are not free agents to do what they would wish to do in the way that trainers and jockeys are.

I welcome the feedback from you both. Keep it coming: perhaps between us we could nail this!

Simon Rowlands | 24 September 2008

Stu,
You'll find plenty of articles about strategies, pricing markets and the concept of value on the Education site - betting.betfair.com/education.

Betfair Education | 24 September 2008

Cheers for getting back to me Simon. I read your other articles and I'm still slightly in the dark about the transition from looking at ratings and then converting it into a tissue price.

My problem is this. I want to be as objective as possible in converting my data sources into tissue prices. For example, lets assume I hypothetically use official ratings for a race. On Friday there is a listed race at Ascot. Ronaldsay is rated 105, Sell Out 109 and Hobby 101. Presuming Sell Out's mark is accurate, all other things being equal, he should win.

So, is the next step just subjectively arriving at a % figure on what I think his chances of winning are? It's the moving from data sources to this judgement call on what I think the % chance of winning that is the problem for me.

Cheers,

Stu

Stu | 25 September 2008

Sorry for the delay in replying: I have been away for a couple of days.

Here are some of the reasons why I cannot give a simple answer to the problem:

1) It depends on the accuracy of the rating system being used.
2) A single rating does not necessarily encapsulate the horse's likelihood of running to a rating, just that it is, theoretically, capable of running to that rating.
3) The significance of a rating in relative terms varies according to the specific type of race a horse is running in.

Where 2) is concerned, you ideally need some sort of measure of the likelihood of a horse running to the rating, exceeding the rating or dropping under the rating, which will be a function of consistency and potential besides anything else. I know of someone who has used ratings converted into odds for a long time and I believe his model calculates a theoretical % likelihood of a 109 horse running to 109, to 108, 107 etc and to 110 or more. You then match this against the % likelihoods of all horses in the race running to various ratings to come up with probabilities for outcomes, win and place.

Where 3) is concerned, you have an issue in things like novice chases where there is a significant possibility that a horse may run badly/fail to complete, and ratings need to be viewed in that context. At the other end of the scale, the rating of a horse in a Group race on the Flat often reflects accurately both what that horse is capable of and what it is likely to produce. The likelihood of bombing out entirely is also often reduced greatly. In such circumstances, a point or 2 of ratings can be worth a big gap in percentage chance of winning.

That, and the fact that I don't currently do my own ratings for all racing, means that I still stick with subjectivity.

Nonetheless, when looking into this previously, I have attempted to identify those occasions when ratings seem to have done the job of pricing up accurately and to establish a few rules around that. You may want to do likewise.

In an example like the one you gave, the "true" answer may be something like the following:

Ratings: 109/105/101, average 105
Percentages: 100/3, average 33.3%

105 rating nominally equals 33.3%.

109 rating=33.3%*1.11=37.0%, 37.0%*1.11=41.0%, 41.0%*1.11=45.5%, 45.5%*1.11=50.6%

101 rating=33.3%/1.11=30.0%, 30.0%/1.11=27.0%, 27.0%/1.11=24.3%, 24.3%/1.11=21.9%

(the number of iterations depends on the number of points the rating differs from the average).

50.6 + 33.3 + 21.9 = 105.8%, bring this back to 100% and you get:

Sell Out 47.8% (2.10 Betfair odds)
Ronaldsay 31.5% (3.20 Betfair odds)
Hobby 20.7% (4.80 Betfair odds)

I suggest playing around with this process, or a similar one, and with the multiplier/divider being used (1.11 in this instance) until you are happy that it has some application to the ratings that you are using, while remembering the remarks made previously about the context in which the ratings are found.

I intend doing just that myself when I can find the time!

Simon

Simon Rowlands | 27 September 2008

Just a bit of clarification about that process. I knocked up a spreadsheet along the following lines:

Column A: Ratings
Column B: Ratings compared to average rating
Column C: = $A$22*(POWER($A$24, B3), where A 22 is a fixed cell of 100/field size and A24 is a fixed cell defining the "to the power" chosen, e.g. 1.13
Column D: %s brought back to 100
Column E: %s converted to decimal odds

E.g. 4:25 at Ascot today:
(adjusted Racing Post ratings, 1.11 power):
DISTINCTION(124) 7.0
TUNGSTEN STRIKE(121) 9.4
BALKAN KNIGHT(124) 7.0
LAND 'N STARS(111) 26
METAPHORIC(122) 8.6
PETARA BAY(123) 7.8
SAGARA(124) 7.0
SANBUCH(126) 5.6

(adjusted Timeform ratings, 1.15 power)
DISTINCTION(126) 5.1
TUNGSTEN STRIKE(121) 11
BALKAN KNIGHT(123) 8.0
LAND 'N STARS(114?) 30
METAPHORIC(122) 9.2
PETARA BAY(122) 9.2
SAGARA(124+) 7.0
SANBUCH(126) 5.1

I suggest some allowance is made for symbols attached to Timeform ratings, e.g. +6 for p, +12 for P, -6 for ?, +3 for +, which would give:

DISTINCTION(126) 5.5
TUNGSTEN STRIKE(121) 11
BALKAN KNIGHT(123) 8.4
LAND 'N STARS(114?) 70
METAPHORIC(122) 9.6
PETARA BAY(122) 9.6
SAGARA(124+) 4.8
SANBUCH(126) 5.5

I would just like to emphasise again that I have generally priced up subjectively and that this is "thinking aloud" for me, though it does seem as if it represents a workable starting point in the examples I have looked into.

I hope to pick the brains of some who have much more experience in this area than have I in due course.

Simon

Simon Rowlands | 28 September 2008

Wow, thanks Simon! I'm going to have to get a cuppa and a piece of paper to work through this, but I'm sure it will be mega useful, not least of all in making me think a bit more...

Stu | 29 September 2008