Silly Question Friday: Part 2

Welcome to Part 2 of Silly Question Friday. If you missed Part 1, it's well worth a look and can be found here. In these Friday posts, I will respond to those long-term unanswered irritations you might have.

Eyes down, look in, then...

 

What do the values in IV and A/E actually stand for and what are good or bad values?

From: Paul S

What do the values in IV and A/E actually stand for and what are good or bad values?

Paul

Geegeez writes...

This is a good starter for part 2 of SQF, because these numbers are everywhere across geegeez.co.uk and I'd like to have a reference point for people to read and understand why we use them, and how they can help your betting.

In the simplest terms, IV - Impact Value - is a measure of how often something happens in a given situation compared to all of the times it happens. For instance, how often a jockey wins races compared to how often all jockeys win races.

A/E - Actual vs Expected - is a measure of whether a statistic might be profitable going forwards.

In both cases, a figure of 1.00 is 'par' or standard. A number above 1 is good, a number below 1 is not so good, and the further above or below 1 the number, the better or worse it is. Thus, an A/E of 0.4 is likely to be extremely damaging to one's bankroll over time, whereas an A/E of 1.4 would be an exciting find (if on a vaguely meaningful sample size and with logic to support the statistic in question).

So that's what IV and A/E stand for, but how are they calculated?

Let me start by saying that you absolutely do not need to know this: if you only know that 1 is standard, more than 1 is good, less than 1 is less good, and the further away from 1 the better or worse a statistic is... then you know all you need to know.

But, for the curious, here are how the numbers are arrived at...

How to Calculate Impact Value

IV is slightly easier to calculate, as follows. Let's say we want to know how often Mark Johnston wins with his 2yo first time starters at Goodwood compared with the average win strike rate for Goodwood 2yo first time starters overall.

First, we need to know Johnston's record which, for the last five years at time of writing, is six wins from 30 such runners.

Next, we need know the runners, winners and, therefore, strike rate, of all such runners in the same time frame. Those figures are 30 winners from 367 runners, 8.17%

The formula for Impact Value is

IV = %age of winners fitting criteria / % of runners fitting same criteria

In our example, that means %age of Goodwood 2yo 1st time 5 year winners trained by M Johnston / %age of Goodwood 2yo 1st time 5 year runners trained by M Johnston

 

The first bit, %age of winners fitting the criteria, is 6 MJ winners / 30 all such winners = 20%

[Mark Johnston has trained 20% of the 1st time starter Goodwood 2yo winners in the last five years]

The second bit, % of runners fitting the criteria, is 30 MJ runners / 367 all such runners = 8.17%

[Mark Johnston has trained 8.17% of the 1st time starter Goodwood 2yo runners in the last five years]

 

Therefore, the IV for Mark Johnston-trained 2yo 1st time starters at Goodwood is

20 / 8.17 = 2.45

Mark Johnston is nearly two-and-a-half times (2.45x) more likely to have a 2yo first time starter winner at Goodwood than par. And that is IV, a measure of peer-contextual probability.

 

How to Calculate Actual vs Expected

While IV tells us whether we are more or less likely to get a return from a given approach, it doesn't do anything to help us understand whether the long-term returns from said approach will be positive or negative. Clearly, betting on horses is about both staying in the game (backing an 'acceptable' number of winners), and trying to make a profit (backing horses at acceptable prices). This is where A/E comes in.

Using our Johnston 2yo 1st timer at Goodwood scenario, to calculate A/E we first need to know the actual number of Johnston's winners which, in this example, is six.

Next, we need know the expected number of winners. Wait? What?!

To do this, we use a simple formula based on the starting price (you could just as easily use Betfair Starting Price or even tote return if you were sufficiently minded - here I've used SP which is the A/E to which all geegeez quotes refers), thus:

 

Actual number of winners / Sum of ALL [entity] runners' SP's (in percentage terms)

which we know at this stage to be 6/ Sum of ALL [entity] runners' SP's (in percentage terms)

 

To establish a runner's SP in percentage terms, we do the sum 1/(SP + 1).

For instance, 4/1 SP would be 1/(4 + 1), or 1/5, which is 0.20.

And 1/4 SP would be 1/(0.25 + 1), or 1/1.25, which is 0.80.

And so on...

The sum of Johnston's expected 2yo 1st time Goodwood winners, calculated in the above manner, is 4.5312.

 

Thus, Actual / Expected is 6 / 4.5312 = 1.32

 

This is a positive figure.

 

Using IV and A/E in concert

As punters, depending on our attitude to risk and to losing runs, the optimum combination is something which is more likely than 'normal' to happen, and which has a positive value expectation (i.e. is expected to be profitable).

By using these figures and isolating statistics with numbers above 1 for both IV and A/E, we have potentially attractive betting propositions. Of course, we can - and should, where time allows - invest in checking suitability of conditions, pace, and so on, based on form in the book (where there is form in the book).

But these numbers are powerful by themselves in terms of understanding whether something is a quirk, or skewed by a single big-priced winner, or is a little thread of gold to be woven into your betting fabric.

Hopefully that's useful. Next question...

*

Is Richard Johnson’s record of riding odds on chances the worst of all the jump jockeys?

From: ANTHONY K

Your first 30 days for just £1

Matt

Is Richard Johnson’s record of riding odds on chances the worst of all the jump jockeys?

Tony

Geegeez writes...

No, not by a long chalk. When talking about best or worst records, I guess we ought to start with two concepts: winning strike rate, and betting profitability. These can be evaluated with our old mates, IV and A/E respectively. Let me illustrate with the top ten NH jockeys riding odds on shots in UK races in the last five years (as at 25th September 2019 - figures from geegeez's Query Tool)

This first view is sorted by IV - Impact Value - on the right hand side, and we can see that Richard Johnson is in fourth place. It is important to keep in mind that the jockeys at the top, who are four to five times more likely to ride an odds on winner, are also much more likely to ride a horse at 1/6 as opposed to others in the sample whose only exposure to odds on rides is at, say, 4/5 or 10/11. In other words, 'odds on' is a broad church and most of the steering jobs go to the biggest names.

That sort of anomaly gets flushed out when using a market barometer, so let's re-sort the list based on A/E, Actual vs Expected.

For context, I've included the full set of jockeys to have ridden 25+ odds-on shots in UK NH races in the last five years. Johnson is in 11th place, with an A/E as close to 1 as doesn't matter. Backing all of his odds on runners on an exchange or with early BOG prices would have made a profit. Not a life changing one, but then what are you expecting from this approach?

Incidentally, does the fact that Paddy Brennan is bottom of the A/E pile make him a bad jockey? No, NO, NO! He's just not been profitable to follow in this, somewhat contrived, context.

Next!

*

How do you manage to narrow the day's racing down to just one Stat of the Day?

From: Norman A

How,with all the racing we have these days, do you manage to narrow it down to just one stat of the day?

Very successfully I may add…

Norman

Geegeez writes...

Straight over to Chris, aka Mr Stat of the Day, for this one:

So, essentially, here’s how I work…

I have loads of saved angles in Horseracebase and geegeez Query Tool.

That gives me a list of potential qualifiers for the next day.

I then sort them into race order, delete all those in races that I don’t like to get involved in (i.e. novice races on the flat, median auction races, novice hurdles and bumpers).

Then I get rid of any race with over 14 runners, as there's too much scope for hard luck in running.

This leaves me with a more workable shortlist of races.

From there, I go through the races using the geegeez cards & tools to see what I think might win that race. If the horse I like is on my list of angle qualifiers it gets shortlisted as a potential Stat of the Day pick.

I then remove all those priced 5/2 & shorter and finally go with the one I feel has the best chance of winning at a price that can stand a bit of movement. NB: it’s only at this point that I look at prices.

It’s probably a more long-winded process than it needs to be, but as it also generates the Stat Picks selections, it justifies the time.

Chris

*

How are odds for virtual racing calculated?

From: David G

Hi Matt,

Your starter for 10 then (20 if you count it as two questions). How are odds calculated for virtual racing, and when is the result "known"?

David

Geegeez writes...

Not such an 'off track' question as it might first appear.

Odds for virtual events are, in a value sense, irrelevant as these are games of pure chance. There is no form to consider, there is no means of gaining an edge. The result is determined by a random number generator loaded based on the odds of the runners in the race.

As an example, a 3/1 favourite (theoretical 25% chance before overround is applied) might be given a 20% chance of winning by the number generator. A 9/1 shot (theoretical 10% chance) might be given an 8% chance of winning, and so on. The difference between the chance given by the algorithm and the odds available is bookmakers' margin. In the same way that a casino may lose on any single spin or any single day on the roulette tables, they know that their edge over time makes virtual a guaranteed profit product.

The result is known instantly because virtual racing is unencumbered by such things as stewards' enquiries, weighing in light, taking the wrong course and so on.

If you like betting games like roulette, you might as well have a crack at virtual racing. But if you prefer to put past history and racing statistics to work for you, keep it geegeez and dodge the virtual!

*

How does a commentator know whether a horse is on or off the bridle?

From: Michael K

How does a commentator know whether a horse is on or off the bridle?

Michael

Geegeez writes...

If a jockey is sitting quietly on a horse, without getting lower in the saddle or exercising his arms, that horse is on the bridle.

If a jockey has his head and torso closer to the horse and is moving his arms to shake the reins a little, the horse is coming off the bridle.

By the time horses are contesting a finish, normally all horses will be firmly off the bridle and the jockeys will be working hard to encourage them to run as fast as they can.

It is worth remembering that being on or off the bridle is not necessarily a good or bad thing: some horses come off the bridle early habitually but find plenty for pressure and still win races, whereas others - sometimes called 'bridle horses' - travel beautifully through their races but, when it comes to putting their head in front at the business end, they find very little for their rider's encouragement.

*

How long should I persevere with a tipster before throwing in the towel?

From: Andrew F

Hi Matt

Several months ago geegeez recommended a service so naturally I researched it and subsequently subscribed to it.

The first month was great then three disastrous months! I decided enough was enough so I wrapped it in. This seems to be the general pattern for me when subscribing to a tipster, so my question is how long should I persevere with a tipster before throwing the towel in bearing in mind I always research the last 12 months results before I sign up to anyone.

Many thanks

Andy

Geegeez writes...

A GREAT question!

So here's how this works: nothing goes up and up and up. We all know that, right?

And past performance is the best barometer of future performance, without being a guarantee of replication. We all know that, too.

So it can be that something which has fared very well historically is unable to replicate that success.

Naturally, the more evidence we have in the past, the more confident we can be about the future. Not certain, but more confident.

What the past is especially good at, when there's a body of evidence, is telling us about the patterns we can expect. It's not surprising, given how sales copy tends to be, that we're invited to focus on the positive; but good investment discipline is about always understanding the negative.

Following a system or service is a form of investment, and should be treated as such.

That means a betting bank. A separate betting bank.

It means disciplined staking.

It means awareness of potential drawdowns (losing runs).

It means retaining focus and discipline during those drawdowns - and during upswings, too.

In other words, it is not for everyone. In fact, such an approach is probably not for most of the people it attracts.

In my view, you've done everything right before getting involved, specifically a) learning about a service through a trusted portal (geegeez), and b) researching the long-term results set.

But in the execution, is it possible that you've not set up your betting bank appropriately? I'm assuming that the drawdown is in line with what might have been expected based on the previous year's results. (Incidentally, a year may not be enough: it very much depends on how many selections are in that sample, what average odds, and so on).

If this is an outlying drawdown - in other words, if it is worse than has been the case in recent memory - then you should expect some commentary about that from the service provider. But if it is not out of the ordinary, it is simply one of the things that happen on the way to profit. And those who don't have an investment mindset will draw stumps before the next upswing.

One other point about sales copy for betting services - all investment services in fact, just look at unit trusts and the like! - is that they ALWAYS pitch on the back of good recent results.

Again, as everyone knows, after a good run, what should we expect? A bad run. And, if a service is proven and decent (think Stat of the Day), what should we expect after a bad run? You're ahead of me now 😉

I've tried to keep this reply generic because it is an excellent question and something with which a lot of people struggle. I hope I've touched on some of the possible reasons why.

Mindset, primarily.

But also, and again, this can easily happen in regulated markets like share trading, forex and unit trusts, sometimes services which have performed well historically simply fail to replicate that performance in the future. Nothing is guaranteed. Nothing is risk-free. But setting up with an investment mindset, undertaking due diligence first (as you have done), and executing your investments stoically and dispassionately even if it means going deep into your betting bank... this is the way to give yourself every chance of success.

I do not follow services, because this approach doesn't work for me. I love the puzzle. I want to find my own answers. And I want action bets along the way. Regular small bets for action, less frequent bigger bets for fancies. Ultimately, I probably don't have the discipline for following services, and in any case it doesn't 'scratch my itch', which is to engage with the puzzle and, to some degree, 'to be right'.

Everyone bets for different reasons, and understanding a little about our motivations - in the context of what is possible, and what is required from a given approach - gives us the best possible chance of satisfying those motivations, whether they are to pass time, to solve the puzzle, to be right, and/or to make some money.

Phew, hopefully the above makes sense and offers something upon which to mull.

*

Have you done any work/given advice on bankroll management?

From: Phil M

Hi Matt,

Have you done any work/given advice on bankroll management? From how to build a bankroll from scratch to what to do when you have one... working out whether a system/method is scalable?

Cheers

Phil

Geegeez writes...

The short answer is 'no'. I don't generally like to offer advice on bankroll management, but on this occasion I will elaborate a little.

The reason I tend to steer clear of such advice is that it borders on financial advice, for which I'm not regulated. Moreover, such advice differs from person to person based on attitude to risk, available funds, discipline/mindset, betting approach, average odds, backing/laying/dutching, and so on.

Staking advice is a step further into the unknown, and the general principle here is that, if it isn't profitable to level stakes, bin it. Trying to get creative with staking is generally akin to trying make a silk purse from a sow's ear.

How to build a bankroll is an interesting question, and it involves discipline and small incremental gains. The simple fact is that if you have a 'short stack' to start with, it will take some time to get to a place where you have a bankroll more in line with what you'd like. Again, this taps in to discipline and money management and, further, it begs the fundamental question: why am I betting?

If betting for fun as well as profit, you're in good shape building a small bankroll into a larger one, because the fun element will nourish you when the incremental profit gains are small.

If profit is a primary motive, I don't really feel qualified to comment. I'm not a professional investor. I do bet with a profit expectation, but I bet primarily for fun. The two, as I never tire of saying, are NOT mutually exclusive. Betting primarily for profit may involve the dark arts of bad each way, bonus abuse, and any number of routes to profit which are not really related to the enjoyment of betting. Like I say, I don't feel qualified to comment on such approaches.

The above is more than I generally offer on this subject, though perhaps less than you might have hoped for. I hope at least you understand my reasons for falling short with this particular reply.

*

Whatever happened to Nick Mordin?

From: Terry B

Hi Matt,

This has been bugging me for a couple of years now - whatever happened to Nick Mordin?

Is he still 'involved'?

Cheers

Terry

Geegeez writes...

I'm given to understand Nick Mordin has retired. In truth, he's not been around for quite a long time. Last I knew, he was writing a weekly piece for the Irish Field and, before that, the Weekender. I heard a rumour that he was advising some larger gambling syndicates and had spent some time in north America (New York, I think). But I really don't have anything but hearsay and conjecture to add, which is to say I don't have anything to add.

*

Which UK horses have a chance in the Melbourne Cup?

From: Graham F

Hi Matt

Which UK horses have a chance in the Melbourne Cup? A friend of mine thought Basanti might be a rough chance. See it is running at Doncaster tonight.

Cheers

Graham

Geegeez writes...

Not really an area on which I focus. The main reasons are that a) the Melbourne Cup comes hard on the heels of the Breeders' Cup, into which I pour a lot of energy and b) I don't have any handle on the Oz form. Also, c) the final field emerges very late in the day with qualifying races happening as close as the week before the Cup.

That said, the record of Euro, or ex-pat Euro, horses is exceptional. Cross Counter last year led home a British 123, with Europeans also filling out 5th, 8th, 9th, 11th and 12th in a field of 24. Only two Euros finished in the second half of the field and one of those pulled up lame.

In 2017, Rekindling, trained by Joseph O'Brien, prevailed. It was another 1235, with Euros also in 6th, 7th, 9th, 10th, and 11th places.

And in 2016, the ex-French Almandin just edged Irish-trained Heartbreak City into second, with ex-British Hartnell third and Charlie Appleby's Qewy in fourth.

The problem, as you've hinted at in your question, is that a lot of Euros head south. However, with average winning odds of 9.7/1 from the five Euro victors since 2010, you can afford to back three or four each way and still hope for a return. Good luck!

*

Are there any good tipsters who give tips out in the morning, and who make a profit at BFSP?

From: Ian B

Are there any tipsters, maybe with your good selves who give tips out am, not the night before which is hopeless, and who make a profit at bfsp?

Ian

Geegeez writes...

There are a number of good tipsters, most of whom - as you say - send out selections the night before.

One notable exception is Hugh Taylor on the ATR website. His picks are usually online around 10am and, while the traditional bookie prices go quickly, I'm given to understand he shows a small profit at BFSP.

Of the reviews we've undertaken, those which were favourable can easily be viewed here and, by clicking into them, readers may understand about their suitability for their individual betting style. (N.B. an update to that table is overdue, and will be published imminently).

**

That's all for this first edition of SQF. I hope you now know at least one thing you didn't when you started reading this! If you've any more questions on racing, betting, Geegeez Gold, and so on, please do drop me a line and I'll add it to the list!

Matt

10 replies
  1. Eddie
    Eddie says:

    Hi Matt,

    Was hoping my silly question would come up so I’ll take my chances here and hope you’ll respond!

    When looking at flat racing, whether AW or Turf, how do you determine which stalls are low, med, and high? For instance if it’s 9 runners with 9 stalls, that be easier where you can say 1-3 low, 4-6 med and 7-9 high, but what about when there’s an odd number of runners? Eg 13 or even in the big races when there’s 20+ runners or even at the other end when there’s like 5 runners?

    Thanks

    Eddie

    • Matt Bisogno
      Matt Bisogno says:

      Ran out of room, Eddie. It’s 4000 words already!

      Will definitely cover your question next time. Suffice it to say, 10 runners is 343, so you can probably work out the algo from that.

      More next time!

      Matt

  2. TheLoneBettor
    TheLoneBettor says:

    Hi Matt,

    My “silly question Friday”, well maybe not silly at all but just curious.

    Are your ROI figures a combination of both Win and E/W bets or do you just calculate Win bets only?

    Cheers

    Shane

    • Matt Bisogno
      Matt Bisogno says:

      Hi Shane

      ROI is calculated on the win part only, unless explicitly stated.

      Matt

      • TheLoneBettor
        TheLoneBettor says:

        Brilliant. That’s has put my mind at ease. And this “Silly Question Friday” is also brilliant.

  3. Kevin
    Kevin says:

    Matt,

    my ‘silly’ question – is there a website that has registered all racing horses? i ask this as there are a couple in my tracker that havent raced for over a year and I have no idea whether they are still active in racing or not !

    • Matt Bisogno
      Matt Bisogno says:

      Hi Kevin

      The BHA keeps a record of ‘horses in training’.

      You can search their records here: https://www.britishhorseracing.com/racing/horses/

      If your horse is not on that list it is probably out of training, or moved to another jurisdiction. It may of course come back into training, having gone point to pointing, say, so it’s not straightforward.

      Matt

  4. TheLoneBettor
    TheLoneBettor says:

    Hi Matt,

    Loving using Geegeez Gold. I have set up many Angles and for obvious reasons I am paper trading at the moment. The results are fantastic but I’m not getting carried away by it. More research needs to be done though but I love digging down and I know where to stop with the “Back-Fitting”, or so I think.

    Anyhow, my SQF. What is the difference between a Tricast and a Trifecta?

    Now, I have tried to Google this but as with some Google answers they were not clear enough. The reason why I ask is because one of my Angles (Which I have called AW H’cap 1m’s) brought up 3 qualifiers in the same race. The race itself was the 19.00 @ Kempton Park on the 10th October 2019. My 3 Angles were all trained by S C Williams.

    The horses were 1st VIA SERENDIPITY@16/1, 2nd PINNATA@7/1, and 3rd PACTOLUS@33/1

    And yes, that is the order they finished. Bloody paper trading lol. The Tricast paid £3,686.13 and the Trifecta paid £1,808.90 – I do plan to have small stakes on these Angles some time in the future. I know that these kind of bets are as rare as the payout price suggests but with Query Tool I believe you give yourself a better edge. In reality I probably would have done a reverse Tricast (Would have paid £613.35 to £1 x 6 bets) and when such occasions where you have 2 Angles a reverse forecast. The payout difference between the Tricast and Trifecta is just over 100%.

    So Matt, what is the difference between the 2 of them and have you any thoughts on which bet would be best to use?

    Regards

    Shane

    • Matt Bisogno
      Matt Bisogno says:

      Hi Shane

      The difference is that one (trifecta) is a tote bet and the other (tricast) is computer calculated.

      That means the trifecta could have a broader range of dividends, depending on how many other players have the winning sequence.
      The tricast is based on a formula according to price and various other factors, regardless of the number of players with the winning sequence.

      Trifecta is generally better when perming outsiders, tricast when wagering near the top of the market. Generally, but not always: that’s the chance you take!

      Matt

Comments are closed.