Oh sorry.
I thought we were all on here to learn things that might be some PRACTICAL assistance to our gambling efforts,
My apologies.
I just hate time wasters and ****s.
Btw they never make any money either.
So think about it, if you're seriously thinking of making money out of gambling.
Why don't you try to respond to the points I'm making in a constructive manner, if you're really interested in trying to make some sort of useful point on gambling ?.
Not being nasty here, but I am finding this all very dishearteningly pointless.

oky no i got it

FAFH:

I've been as constructive as possible. And I hope being able to calculate your edge is practical.

Please if there's anything you don't understand in my posts don't feel bad about asking for clarification, I often garble off into nonsense (both to a layman and to a maths geek), my communication skills are poor.

I'll work that example again.

Stake 1@2.1, bet loses, lose 1.
Stake 1@2.1, bet wins, gain 1.1*0.95 (commission).

Net profit (total profit from bets won - losses): 0.045.
turnover (total profit from bets won): 1.045

Edge = net profit / gross profit = 4.3%.

Practically you'll have a lot more samples.

Knowing you have a 4.3% edge with 5% commission is useful. Not only that but if you record all your bets individually you could create a spreadsheet where you look to see if you still make money at 6%, 7% commission. This is practical if betfair change their terms.

FAFH: Doing it how kelly did requires knowing the true odds of a future event. You cannot possibly know that, so the best way to try to estimate that is to take your edge from all your previous bets and estimate that this is likely your edge for your current bet.


Otherwise you're just guessing. "Ooh this is at 1.9, I reckon it's probably a 1.8 shot".

That's a guess.

"It's at 1.9, and my selection is made similar to my history of selecting, and my history shows I've roughly a 4% edge, so true odds should be closer to ..." is more practical, and probably more profitable long term.

ror
I have no problem in understanding your calculations.
I just have no use for them, and can see no use for them for others in the real world of gambling.
But if it pulls your chain to fiddle around endlessly thinking about these things, instead of getting out there and dfoing some gambling, then go for it.
It's a free world.
I have spent years surrounding by extremely intelligent boffins ( yes at one stage I did work in finance) and I came away totally and utterly unimpressed by either their common sense, or more worryingly their actual bottom line contribution.

Not his last post, the one a few before.
Whatever.

IronMan:

It doesn't look like you're doing anything wrong.

In your backing example your edge (naively calculated without details on every bet) is:

strikerate-1/(average odds)

(Assuming these are decimal odds so 2.0 would be evens)

This equals:

(412 / (412 + 1 304)) - (1 / 5.076101399) = 0.0430916593

So that's a 4.3% edge, which is huge. (Roulette has a 3.2% edge and the house cleans up).

The laying example is more complicated, because you'd need to convert your lay odds into an equivalent ("backing to lose") odds.

I'm guessing here, but does that -0.9836 mean that on average you were laying 1.98 shots?

If that were the case, your edge in here in the laying case would be 3.5%.

As I said upthread, it's probably useful to put confidence intervals around that.

Thanks ror.....and YES  -0.9836 is BF odds 1.9836......

It may be "naive" ....but ROI has to be the strong indicator of whether you have an edge or not.....but thanks for converting it to "edge" mathematically.

On BACKING STATS......AVERAGE ODDSPER BET    5.076101399........is at BF ODDS 6.07 etc

DFCIRONMAN: I would argue ROI is more important than edge, because it takes into account both edge and staking.

henok: That does seem to rather undermine my earlier posts.

"True" edge (rather than just the estimates we are calculating) actually doesn't take into account how much you stake, unless you consider that part of your edge (i.e. you consider that part of what makes you a winner.) therefore shouldn't really be taking into account money won/lost, but if you're level staking then the two shouldn't be far off each other.

The ROI takes into account commission tho which I realise I forgot to do for IronMan, but a 4% edge level staking is more than enough to withstand a 5% commission.

BTO if you think your profit is all the winning bets you've made without taking into account all the losing bets you've ever made, no wonder you "beat the overround".

Also in the first case,
a) I didn't use it because I don't like the definition.
B) You say I'm wrong then produce exactly the same definition! i.e. total amount staked. Amount staked IS your liability on each bet.

AFTER

Ror, I never said profit is all the winning bets you had.
You said that.
There is turnover.
There is profit.
There is profit on turnover.

They are seperate entities.

Your winning are your profit after deducting
your total turnover.

That's cash outlayed to gain the profit.
Nothing to do with winning bets included.

For clarity:

Total stakes
Total return
Total profit
Profit on turnover

Liability doesn't come into it, because you already
included it in the return if it happened.
If it didn't happen, it shouldn't be included.
Just as the potential 1000/1 winner never happened
that you could have claimed!

BTO:

Profit is Income - Expenses.

Income is all your winning bets.
Expenses is all your losing bets.

That Income is all the profit from each of your winning bets.
That Expenses is the stakes of all your losing bets.

If you're getting confused about using profit twice, you could think of it as gross profit and net profit, although that's not strictly the case.

I'm saying turnover should be that income.

Put it this way, if you put a tenner on at 5/1, what do you consider your turnover to be, what do you consider your income to be?

Would you put a tenner into expenses and 60 into profit, or would you simply stick 50 into profit?

I think here we have a clash of old school (odds and "returns" etc) and new school (decimals, profit vs loss).

If you want you can count turnover as all the stakes you've made, both winning and losing, but if you do do that you should count returned stakes as well as profit in your ROI.


In other words, you place two bets at 5/1, one wins, one loses.


In my world, your income is 5, your profit is 4, your expenses is 1.

In your world your income is 6, your profit is 4 and your expenses is 2.

I can see arguments for calculating it either way, which is why I gave both options in my original "what is turnover" post.

As for laying, well personally I see laying as backing to lose, so it's just backing and can be calculated the same way, you just need to formulate the equivalent "to lose" odds.

I agree but we need clarification: If I bet a quid at 2.1, then 1.1 is my profit, 2.1 is my return in a betting shop sense, but 1.1 is my return in a "return on investment" sense.

If you lay 1 horse to a liability of 100 then you can lose 100.

If you back multiple horses or lay multiple horses then it gets far more complicated, in which case you probably need to just consider max liability.

How are losing bets included in your income?

As I said, I can see the appeal of both methods of calculating ROI.

You could say that if you had £1000, split it 100 ways, went out put 100 bets on, then looked at your returns from those bets then that is the ROI.

And that works well, but then you surely need to take the stakes into your returns.

Personally I see it as a more organic profit/cost. There's no chance you'll lose all 1000 bets so really the 'investment' is how much money you need to spend on losing bets to generate the winning ones.

I realise I'm probably going against convention here, which is again why I provided my first defintion of turnover which is all your liabilties (stakes).

So yeah, use profit/stakes but personally I think that is old fashioned and you can be more aggressive about what your ROI is.