BR - management depending on ROI

[Stin]

Let’s imagine 2 professionals.
1st make 1000bets/month with ROI~2%.
2nd make 40bets/month with ROI ~20%.
Average odds for both 2.5(+150 US).

Both want to use flat system.
What unit size in % do u recommend for each one?
Any calculation ideas?

[veriableodds]

Let’s imagine 2 professionals. -- lets assume both have same starting bankroll of $100
1st make 1000bets/month with ROI~2%. --- risk of 0.10 per play or 1/10%
2nd make 40bets/month with ROI ~20%.--- risk of 2.5 per play or 2.5%
Average odds for both 2.5(+150 US).


starting numbers doesn't answer any questions.

[Stin]

So, u think 0,1% and 2,5% is optimal? Because using of 100% roll/month?

i don’t really think number of bets is a key factor here. We could change it to 500/100, so 0,2% / 1% then?

starting numbers doesn't answer any questions.

What does it mean?
higher ROI -> lower risk, lower ROI -> higher risk. If it’s true, we should use different unit size or not?

[Stin]

lets assume both have same starting bankroll of $100

This doesn’t really matter also.
But ok, both have 50000$, betting on a markets with 3000$ max possible bet, both use Pinnacle for betting.

[veriableodds]

I would like to hear others opinion on this, for me you have valid points, and a great start for formatting your own formula . Its all based off bottom line example
------
starting bank of 5,000
total wagers 3,000
risk of 0.3 (15)per wager (0.3 of 5k)
money laid 45,000
roi is 8%
profit would be 8% of money laid= 3,600

[TommieGunshot]

First one should be fine at 1% of bankroll.

Second one should be ok at 5-10% of bankroll. On that, I would go with the high end when there is only one unresolved bet and the low end when there is more than one -- at only 40 per month shouldn't ever be much more than two. Perhaps aim to never have more than 20% at risk at once.

This assumes there are no significant deviations from the average odds. If there are, should use a "to win" amount, rather than an "at risk" amount.

Here is most of the math I used: http://www.herrold.com/brokerage/kelly.pdf

If anyone thinks I applied that incorrectly, please show me where we went differently.

[Gaze73]

First one should be fine at 1% of bankroll.

Second one should be ok at 5-10% of bankroll. On that, I would go with the high end when there is only one unresolved bet and the low end when there is more than one -- at only 40 per month shouldn't ever be much more than two. Perhaps aim to never have more than 20% at risk at once.

This assumes there are no significant deviations from the average odds. If there are, should use a "to win" amount, rather than an "at risk" amount.

Here is most of the math I used: http://www.herrold.com/brokerage/kelly.pdf

If anyone thinks I applied that incorrectly, please show me where we went differently.

10% staking is insane at those odds, wipe out guaranteed. I'd say 3% for the second one.

[TommieGunshot]

10% staking is insane at those odds, wipe out guaranteed. I'd say 3% for the second one.

Full Kelly is 13%. I would bet 100% of my bankroll a player would not get “wiped out” betting anything less than that on bets with 20% edges. How much of that action do you want?

A dealer flashing a hole card in blackjack is ‘only’ a 12% edge. To get 20% in sports would mean getting +140 on a point spread where both sides are -110. If I ever saw that I would bet whatever the max is without question. Now imagine getting to bet on something like that 40 times per month

[Gaze73]

Full Kelly is 13%. I would bet 100% of my bankroll a player would not get “wiped out” betting anything less than that on bets with 20% edges. How much of that action do you want?

A dealer flashing a hole card in blackjack is ‘only’ a 12% edge. To get 20% in sports would mean getting +140 on a point spread where both sides are -110. If I ever saw that I would bet whatever the max is without question. Now imagine getting to bet on something like that 40 times per month

So 10% Staking gives you an over 87% chance to lose 75% of your BR within 1000 bets, are you comfortable with that? 5% stakes reduce it to around 4% chance and 4% stakes reduce it to around half a percent chance.

[Gaze73]

Full Kelly almost guarantees losing half of your BR within just 100 bets, it's for crazy people and robots.

[TommieGunshot]

Full Kelly almost guarantees losing half of your BR within just 100 bets, it's for crazy people and robots.

Bad math. Win is correct at 1.4. Loss should be 1.0. Now see what happens

[HeeeHAWWWW]

i don’t really think number of bets is a key factor here. We could change it to 500/100, so 0,2% / 1% then?

Yep. The effect of many simultaneous bets on the optimal stake is dependent on edge - at 2% and say 10-20 bets it doesn't make a great deal of difference.

Doing this in my head because away from my sheets at home, but you'd need 4% and 50 simultaneous to half the optimal stake.

[Gaze73]

Bad math. Win is correct at 1.4. Loss should be 1.0. Now see what happens

It's correct, the $ win loss ratio is 1.4 to 1.0.

[Gaze73]

Bad math. Win is correct at 1.4. Loss should be 1.0. Now see what happens

Here you have a real life example which happened to me within the first 100 bets; a 1-11 streak. At 13% kelly you'd be in a huge hole despite having a big edge.

[Stin]

This assumes there are no significant deviations from the average odds. If there are, should use a "to win" amount, rather than an "at risk" amount.
.

“To win” it’s a bad strategy,imo. Because most likely our edge is higher with higher odds and lower with lower odds. So we basically bet more money with low ROI and less money with big ROI.

[Stin]

My opinion 0,5% for first one and 3% for 2nd.

[TommieGunshot]

It's correct, the $ win loss ratio is 1.4 to 1.0.

I understand that part now. But I also looked up the 'max drawdown'. That means 50% from the peak, not from the starting amount. So start with $10,000, reach a peak of $200,000, then go back down to $100,000 and it still counts as losing half your bankroll, even though it also means winning 10-times your starting bankroll. Yes the swings will be huge when betting more, but those are necessary to maximize growth. And it doesn't mean "wipe-out guaranteed".

[TommieGunshot]

Here you have a real life example which happened to me within the first 100 bets; a 1-11 streak. At 13% kelly you'd be in a huge hole despite having a big edge.

I'm pretty sure those are not your first 100 bets (you're including players who are still teenagers, so either you just started betting, or you were betting on them when they were like toddlers). We know if anyone makes enough bets 1-11 (and worse) streaks will happen. 11-1 streaks will also happen.

I'm also sure you did not have a 20% edge on those. Probably not even 2%. Change the payouts to reflect a 20% edge and you would have done ok over those 36 bets despite seeing such a bad streak.

[Stin]

A dealer flashing a hole card in blackjack is ‘only’ a 12% edge. To get 20% in sports would mean getting +140 on a point spread where both sides are -110. If I ever saw that I would bet whatever the max is without question. Now imagine getting to bet on something like that 40 times per month

We bet 2.5(+150) with 48% winrate, not a big deal, imo. We still could lose 10 time in a row. No point to bet as much as you can.

[Stin]

Another question , what do we consider a bankroll for prof. bettor(betting is the only income)?
My opinion it’s all what we have(money, car, apartments etc.). If you think the same, I don’t think you would bet 10% or even 5% or as much as you can.

But if you think bankroll it’s just our separate gambling money , ofc you could bet 10% or even 20% .

[TommieGunshot]

We bet 2.5(+150) with 48% winrate, not a big deal, imo. We still could lose 10 time in a row. No point to bet as much as you can.

Finding 40 of those bets per month, would mean being able to turn $20,000 into $100,000,000 in less than three years betting only 3% (winning around 690 at +150, while losing only 750). So no matter what you do decide to bet, you're going to run into limits very quickly. In which case we end up exactly where I said before "bet whatever the max is without question".

[Stin]

Finding 40 of those bets per month, would mean being able to turn $20,000 into $100,000,000 in less than three years betting only 3% (winning around 690 at +150, while losing only 750). So no matter what you do decide to bet, you're going to run into limits very quickly. In which case we end up exactly where I said before "bet whatever the max is without question".

That’s a good point, so we don’t need to rush , even with 2-3% bet we reach the maxbet limits quickly, with almost no risk.

[Gaze73]

I'm pretty sure those are not your first 100 bets (you're including players who are still teenagers, so either you just started betting, or you were betting on them when they were like toddlers). We know if anyone makes enough bets 1-11 (and worse) streaks will happen. 11-1 streaks will also happen.
I'm also sure you did not have a 20% edge on those. Probably not even 2%. Change the payouts to reflect a 20% edge and you would have done ok over those 36 bets despite seeing such a bad streak.

You make a lot of baseless assumptions. And what's wrong with betting on younger players?


These are the overall stats of my ATP small dogs of 2019, and it is my first year betting tennis. It doesn't matter what my edge is, but in this sample size it's 37% and yet I went on a 1-11 streak, because variance is a bitch and often the dogs are dogs for a good reason.