Why do increased stakes lower your chance to win?

[semibluff]

Zero. There is no edge on either side of the bet. Kelly is immaterial. These bets aren’t made as a percentage. I have a feeling the silly claim about break even is not going to be supported with anything meaningful.

Make 20 same-stakes bets at +100 on a true odds +100 probability and there's a 17.62% chance you will go 10-10 and break even. There's a 41.19% chance you'll be winning and a 41.19% chance you'll be losing. If you vary your stakes there will be a different break-even point. There may even be a 0% chance of breaking even. However, the chances of being ahead will still exactly match the chances of being behind, as will be the amounts you are ahead or behind by.

If you bet 1% of 100 units and resize units to 1% of the new bankroll total after each result you will end up with 99.99 units after a W-L 2 bet series, 99.99 units after a L-W 2 bet series, 98.01 units after a L-L 2 bet series, and 102.01 units after a W-W 2 bet series. This is not what was said but perhaps what was implied. 3 of the 4 outcomes are below the starting point but each is equally likely and the true net outcome is net zero change from the starting bankroll.

It's just semantics.

[TommieGunshot]

Can you describe the math behind this?

If I bet on a coin flip 100 times, each time my bet is a random amount between 1 and 100, how much has the break even point gone up?

How about if I make those bets one million times?

The break even point will be significantly higher over 100 flips than over 1M flips.

I'll break it down in a little while. In the middle of something now.

Ever get around to breaking this down? I am genuinely curious as to how someone could arrive at this theory

[budwiser]

This is false on the first post.

There is no coin flip spread. Let's say the nba spread is -4. Even if it's close it's not 4. Maybe 4.1348274. No spread in the history of sports has ever been coin flip odds.

It's the juice that gives the books advantage. So even if the spreas isn't perfect, with the margin for error they come out ahead in the long run.

99% of the time a guy who wagers on 4 games a day is a square. In my opinion

[GunShard]

I also seen this strange pattern when I bet. I seem to lose more often on $1,000 bets but seem to win more often on $100 and smaller $20 parlay bets.

[BrickJames]

Ever get around to breaking this down? I am genuinely curious as to how someone could arrive at this theory

I honestly think KVB explains it best.

If you differ your bet size over a sample size it WILL raise your break even point.

[Johnnythunder]

I guess it's just me then. Whenever I bet big, suddenly statistics don't matter and I'm much more likely to lose.

Same...it's unreal. 4 or 5 times this year already I had very nice streaks going only to "go for it" with a heavy wager. Loss every time. This goes back years for me.

[BrickJames]

Same...it's unreal. 4 or 5 times this year already I had very nice streaks going only to "go for it" with a heavy wager. Loss every time. This goes back years for me.

You're not more likely to lose it just hurts more and has a bigger effect on your bankroll and that's why you are so affected by it.

It's all in your head

[J. v. Neumann]

Let’s keep it simple and get back to basics. Take a set of 21 bets; give yourself 11 wins and 10 losses. Now take any starting bankroll amount. To keep it simple, bet 5.5% to win 5% for each bet and put the wins and losses in any order you desire.
Now, if you didn’t change the size of your bets, 52.4%, or 11 wins and 10 losses would result in breaking even.

Let me just reproduce this...

So with "bet 5.5% to win 5% for each bet" you mean from a 100 units bankroll I would bet 5.5 units, given odds of 21/11 (=1.91) each win would make 10.5, i.e. 5 units or 5%.

So with 100 units and constant stakes, after 11 bets won and 10 lost we would have 100 + 5.5*(21/11-1)*11 - 5.5*10 = still 100 units (as for the break even at 11 of 21 expected).

But Gunshard’s plan requires changing the size of your bets after each play, changing the breakeven point. [...] Try it again and again. No matter what order you place your 11 wins and 10 losses, after those 21 bets, you always have 97.1% of your roll left.

So each bet I either make 5% of my bankroll or 94.5% of my bankroll remains, right? That means after 11 bets won and 10 lost we would have 100 * 1.05^11 * .945^10 = only 97.12 units remained.

I'm not sure because even though I get the same numbers, for the next example...

In fact, after 110 wins and 100 losses with 5.5% bet each time, you would be down somewhere in the neighborhood of 25% of your starting bankroll.

...I get 100 * 1.05^110 * .945^100 = 75 units remained (25 % lost instead of 25 % of the starting bankroll remained, would make sense as a careless mistake...). So did I correctly reproduce your example anyway?

[Gaze73]

Lmao, I thought this thread would be dead on day 1. I'll have to read through the 3 pages.

Basically my point was that even with zero edge there should be a 50/50 chance that you'll run hot with the bigger stakes, but in reality big stakes lead to big losses most of the time. You always catch the worst of the worst of variance with big bets. You like a soccer team as the best bet of the day? They'll get a red card and a penalty in the 10th minute. You like a basketball team? They'll blow a 20 pt lead in Q4. You like a tennis guy? He'll go 2-0 to 2-3 on sets.

[pologq]

funny cause i feel the same when i increase my best size. rarely does that big bet win. happened game 5 of the nhl finals. then i bet less game 6 on the avs and they win.

[texhooper]

Lmao, I thought this thread would be dead on day 1. I'll have to read through the 3 pages.

Basically my point was that even with zero edge there should be a 50/50 chance that you'll run hot with the bigger stakes, but in reality big stakes lead to big losses most of the time. You always catch the worst of the worst of variance with big bets. You like a soccer team as the best bet of the day? They'll get a red card and a penalty in the 10th minute. You like a basketball team? They'll blow a 20 pt lead in Q4. You like a tennis guy? He'll go 2-0 to 2-3 on sets.

I was with you all along…then it got turned into a legitimate mathematical crisis

[KVB]

...In fact, after 110 wins and 100 losses with 5.5% bet each time, you would be down somewhere in the neighborhood of 25% of your starting bankroll...

...75 units remained (25 % lost instead of 25 % of the starting bankroll remained, would make sense as a careless mistake...). So did I correctly reproduce your example anyway?

Notice I said the bettor would be down 25% of the bankroll, not "down to" 25% of the bankroll.

So you I think you did it right, and there is no mistake on either of our ends.

[TommieGunshot]

with zero edge there should be a 50/50 chance that you'll run hot with the bigger stakes, but in reality big stakes lead to big losses most of the time. You always catch the worst of the worst of variance with big bets.

It is somewhat fascinating how someone could do so much mental gymnastics to reach such a ridiculously low level of logic. If big losses happen "most of the time" and the worst variance "always" happens, then it was not "50/50" and the edge was negative, not "zero".

[Gaze73]

I'll put it another way. You have a 100 unit bankroll and bet on 20 markets, 5 units each. NFL spreads and totals. You're obviously overbetting, but still you shoud expect to win 10 of the bets. In reality, you will lose 14 of the bets, but you will never win 14 bets, even though it's just as likely.

I even ran into this problem with an edge by tailing sharps. Let's say you follow 5 top NFL experts who have a 10% long term ROI. Everyone gives you 4 bets for the night, so you risk your entire BR and yet again you'll lose 12-13 bets but never win 14-15. In 10 years I've had maybe a few nights where I overbet like that and came out with profit. 90% of the time the BR took a 20-50% hit.

[cala56]

Thanks KVB im totally agree with you, you should write a book. You YouTube videos are great also.

[Accrued Interest]

Lmao, I thought this thread would be dead on day 1. I'll have to read through the 3 pages.

Basically my point was that even with zero edge there should be a 50/50 chance that you'll run hot with the bigger stakes, but in reality big stakes lead to big losses most of the time. You always catch the worst of the worst of variance with big bets. You like a soccer team as the best bet of the day? They'll get a red card and a penalty in the 10th minute. You like a basketball team? They'll blow a 20 pt lead in Q4. You like a tennis guy? He'll go 2-0 to 2-3 on sets.

Actually you should be subject to less variance on big bets because variance is part of the calculation.

[TommieGunshot]

I'll put it another way.

You mean another dishonest contradiction

You have a 100 unit bankroll and bet on 20 markets, 5 units each. NFL spreads and totals. You're obviously overbetting, but still you shoud expect to win 10 of the bets. In reality, you will lose 14 of the bets, but you will never win 14 bets, even though it's just as likely.

Would you lay 1000 to 1 against me at least seven out of 20? Why not? You said I "will lose 14". If you wouldn't take my offer it means even you understand that everything you say is just ridiculously stupid bullshit

I even ran into this problem with an edge by tailing sharps. Let's say you follow 5 top NFL experts who have a 10% long term ROI. Everyone gives you 4 bets for the night, so you risk your entire BR and yet again you'll lose 12-13 bets but never win 14-15. In 10 years I've had maybe a few nights where I overbet like that and came out with profit. 90% of the time the BR took a 20-50% hit.

10% ROI is like getting paid +120 on a coin flip. Kelly Criterion says to bet 8.33%. Betting only 5% of my bankroll is not overbetting. Being such a big loser makes it quite clear that the bets did not have a 10% long term ROI. The more you talk, the more and more it becomes clear that everything you say is stupid bullshit.

[Gaze73]

You mean another dishonest contradiction



Would you lay 1000 to 1 against me at least seven out of 20? Why not? You said I "will lose 14". If you wouldn't take my offer it means even you understand that everything you say is just ridiculously stupid bullshit

10% ROI is like getting paid +120 on a coin flip. Kelly Criterion says to bet 8.33%. Betting only 5% of my bankroll is not overbetting. Being such a big loser makes it quite clear that the bets did not have a 10% long term ROI. The more you talk, the more and more it becomes clear that everything you say is stupid bullshit.

Oh really? Show me one pro who risks 8.33% of his bankroll per bet. Everyone says to bet 1%, 2% tops.

> Being such a big loser makes it quite clear that the bets did not have a 10% long term ROI
I said I followed top experts who make a living with betting. No matter what year, no matter what sport, whenever I followed multiple pros each night with big stakes I always lost everything. And even at -5% roi I should occasionally win 13-14/20 but it never happened and the opposite happened many times.