Higher R.O.R. = Better Odds of Winning

LovinItAll

Well-Known Member
(I knew I should have taken more time to present this problem. This isn't about a betting strategy that's intended to be implemented in 'real life'. It's a math problem that addresses the question, "Does a lower risk of ruin mean that one will make more money over an indefinite sample size." I said no. I think I'm right.)

Hey,

I was trying to explain to a friend who was asking me how to lower his session ROR [to make more money]. [I said] that lowering the ROR isn't necessarily the way to make the most money. He disagreed, saying that the lower the ROR, the more profit one would make. I tried to explain it like this:

(One unit = $50)

Over ten session with a +EV of 1%, here is the difference between a lower 'goal setting' ROR versus busting out:

Risk 10 Units To Win 10 Units
ROR = ~45%

Will reach goal 55% of the time: $2,750
Will go broke 45% of the time: $2,250

...for a net theoretical win over 10 sessions of $500

Risk 10 Units To Win 5 Units

Risk of Ruin = ~30%

Will reach goal of winning 5 units 70% of the time: $1,750
Will go broke 30% of the time: $1,500

...for a net theoretical win over 10 sessions of $250

So, the higher risk of ruin is the most profitable, [regardless of how long it may take].

My math is right, right? If not, where did I screw up?
 
Last edited:

tthree

Banned
LovinItAll said:
Hey,

I was trying to explain to a friend who was asking me how to lower his session ROR that lowering the ROR isn't necessarily the way to make the most money. He disagreed, saying that the lower the ROR, the more profit one would make. I tried to explain it like this:

(One unit = $50)

Over ten session with a +EV of 1%, here is the difference between a lower 'goal setting' ROR versus busting out:

Risk 10 Units To Win 10 Units
ROR = ~45%

Will reach goal 55% of the time: $2,750
Will go broke 45% of the time: $2,250

...for a net win theoretical over 10 sessions of $500

Risk 10 Units To Win 5 Units

Risk of Ruin = ~30%

Will reach goal of winning 5 units 70% of the time: $1,750
Will go broke 30% of the time: $1,500

...for a net theoretical win over 10 sessions of $250

So, the higher risk of ruin is the most profitable.

My math is right, right? If not, where did I screw up?
You made a mistake in method. It is going to take you about half the time to win 5 units than 10 units. Increasing or decreasing your bankroll isnt going to magically get you paid more for the same bet. The point of risk of ruin is that you will make money on average at a certain rate. You are likely not to actually be at the average at any time. You will be within a certain range a certain % of the time at any time. You are trying to keep as much of that range above zero so you dont go bust(RoR). RoR can increase from higher variance(more risk) or smaller bankroll and the same bets with the same variance. Higher variance does increase the likelihood you will deviate from the average in EITHER direction. You may make more money you may go bust were you wouldnt have otherwise.

If you have alot of money you dont mind losing choose higher variance. Maybe you will deviate wildly on the good side rather than the bad. If you want to consistently add to your bankroll at shorter time periods lower your variance as much as possible, play with as big a spread as you can without a high RoR(player advantage) or catching heat. If you play with a 5% risk of ruin and continue to resize as you win to continue to have a 5% RoR you will go bust on the average after 20 doubling cycles. It could happen at any time even the first.
 
Last edited:

moo321

Well-Known Member
I really can't make heads or tails of this. To the OP, I think you should read some more books before you risk money on blackjack...
 

LovinItAll

Well-Known Member
moo321 said:
I really can't make heads or tails of this. To the OP, I think you should read some more books before you risk money on blackjack...
I think I have a firm handle on money management, BJ, and many other games, but thanks for the advice.

To my original post, I should have added that a never-ending session bankroll is given, hence my reference to a session bankroll as opposed to managing one's bankroll via Kelly betting or other risk management strategies. I'm asking strictly about the mathematics, not the soundness of an overall strategy, once again referencing the theoretical gain as opposed to considering variance.

Thanks...
 

LovinItAll

Well-Known Member
Perhaps I should clarify: His position was that lowering ROR increased theoretical profits. I said they were not correlated the way he imagined they might be. Does that help?

Either the math is right or it isn't - there aren't any other considerations for this particular problem.
 

LovinItAll

Well-Known Member
tthree said:
You made a mistake in method. It is going to take you about half the time to win 5 units than 10 units. Increasing or decreasing your bankroll isnt going to magically get you paid more for the same bet. The point of risk of ruin is that you will make money on average at a certain rate. You are likely not to actually be at the average at any time. You will be within a certain range a certain % of the time at any time. You are trying to keep as much of that range above zero so you dont go bust(RoR). RoR can increase from higher variance(more risk) or smaller bankroll and the same bets with the same variance. Higher variance does increase the likelihood you will deviate from the average in EITHER direction. You may make more money you may go bust were you wouldnt have otherwise.

If you have alot of money you dont mind losing choose higher variance. Maybe you will deviate wildly on the good side rather than the bad. If you want to consistently add to your bankroll at shorter time periods lower your variance as much as possible, play with as big a spread as you can without a high RoR(player advantage) or catching heat. If you play with a 5% risk of ruin and continue to resize as you win to continue to have a 5% RoR you will go bust on the average after 20 doubling cycles. It could happen at any time even the first.
I understand completely, so please let me clarify that we were discussing ROR as it related to potential theoretical profit dollars without any other consideration. My statement was, "Lowering risk of ruin doesn't mean that one makes more money. In fact, with +EV games, the opposite is actually true." I still stand by that.
 

Sucker

Well-Known Member
LovinItAll said:
"Does a lower risk of ruin mean that one will make more money over an indefinite sample size." I said no.
You are correct.

All other things being equal, reducing your risk will ALWAYS lower your expectation. Where in the world did you ever find someone who didn't understand THIS simple proposition?
 
Last edited:

Midwestern

Well-Known Member
LovinItAll said:
Perhaps I should clarify: His position was that lowering ROR increased theoretical profits. I said they were not correlated the way he imagined they might be. Does that help?

Either the math is right or it isn't - there aren't any other considerations for this particular problem.
Just think about it this way. Your variance will be identical if you have the same bet size and spread.
Ror comes into the equation when you consider that variance as a % of your bankroll. It is much easier to double a bankroll of 500 spreading 10-50 than it is to double a bankroll of 40.000. But it is also much easier for the 500 bankroll to go bust before it doubles.

The ONLY thing that can change profitability of card counting is changing either the size or frequency of bets. Changing the $ amount by which you measure your returns has nothing to do with it and is an arbitrary assumption
 

Midwestern

Well-Known Member
LovinItAll said:
Risk 10 Units To Win 10 Units
ROR = ~45%

Will reach goal 55% of the time: $2,750
Will go broke 45% of the time: $2,250

...for a net theoretical win over 10 sessions of $500

Risk 10 Units To Win 5 Units

Risk of Ruin = ~30%

Will reach goal of winning 5 units 70% of the time: $1,750
Will go broke 30% of the time: $1,500

...for a net theoretical win over 10 sessions of $250

My math is right, right?
LIA--
I have no idea where you're getting these numbers from.
But you and your friend Should check out some statistics books to solve your quandry. Figure out what the standard deviation of your play is in $ per hand (not session) and acquaint yourselves with a normal distribution chart. Figure out what z statistics and confidence intervals are and you should have a better grasp of what ror truly means. It has nothing to do with profitability, and how big you determine your ror to be cannot affect profitability whatsoever.
 

LovinItAll

Well-Known Member
Midwestern said:
LIA--
I have no idea where you're getting these numbers from.
But you and your friend Should check out some statistics books to solve your quandry. Figure out what the standard deviation of your play is in $ per hand (not session) and acquaint yourselves with a normal distribution chart. Figure out what z statistics and confidence intervals are and you should have a better grasp of what ror truly means. It has nothing to do with profitability, and how big you determine your ror to be cannot affect profitability whatsoever.
Oy....

No. Apparently everyone is missing the point entirely. This has nothing to do with SD or normal distribution. This is a volume problem. I never said RoR had anything to do with profitability other than the original statement I made, which was comparing two RoR models and analyzing them for profitablility in dollars, not percentage. One last time:

Assumptions:

1) EV is +1% for both players
2) The sample size is large enough to eliminate variance (infinite, if that's what you want)
3) Both players will play the same number of sessions.

Player A: Stops playing when goal of +5 units is reached (or goes broke)
Player B: Stops playing when goal of +10 units is reached (or goes broke)

Obviously both players will enjoy the same return on their investment, which will be 1%. Both players will not, however, make the same amount of money, as Player B's handle will be much larger than Player A's.

In 'the long run', Player B will have wagered more money than Player A because Player A stops playing sooner than Player B.

Another example:

In BJ, over an infinite number of samples, both players are dealt the same cards. When Player A reaches his goal of +5 units, he walks away. Player B, up the same 5 units, continues to play, still enjoying the same +1% EV. Since he continues to place money on the table, his total handle will be larger than Player A's, so his 1% of the total amount wagered will be larger in dollars.

================================
Since the mathematics seem to be eluding everyone, I'll ask this:

Which would you rather have?

1% of $5,000,000 or 1% of $10,000,000?
================================

I think the good people answering this thread are injecting variables that don't exist for this problem. For instance, this is not a gaming strategy issue (unless you have an unlimited bankroll and place no value on your time), nor do standard deviation or any other statistical measure matter in any way. As I said, both Player A and Player B enjoy an EV of +1%. Whichever can get the most money on the table over whatever amount of time it takes to render variance a non-factor will win more money in dollars.

==========================================

And finally....

I'm not confusing RoR with anything, but my friend is. He thinks that the lower the RoR the more profitable the outcome. That isn't true, of course. In reality, RoR is not a factor in profitability unless one is comparing two RoR models as I have.

Seriously, taking my initial statement at face value cannot yield any result other than the one I originally stated. Again, Player B will simply get more money on the table than Player A. Since they will both will be returned 1% of the total amount wagered, Player B will walk away with more money than Player A at the end of the sample.

If that's not clear, I'll be more than happy to provide a mathematical proof, but the problem is so intuitive that I can't imagine anyone needing it.

Best ~ LIA
 

LovinItAll

Well-Known Member
Sucker said:
You are correct.

All other things being equal, reducing your risk will ALWAYS lower your expectation. Where in the world did you ever find someone who didn't understand THIS simple proposition?
YES!!!!!

Hey, I didn't say my buddy was the brightest....

Edit: I'll cop to a bad title for this thread - it was mis-worded and not intentional.
 
Last edited:

tthree

Banned
Take your a example to an extreme and you will see the flaw in your logic. Make player B's goal be 10000. He will be playing many many days to win a session successfully and lose most sessions eventually if the RoR is high enough. Is player A going to stop gambling until B decides his session. I dont think by breaking it into sessions rather than time spent at the table is a legitimate comparison. If they use the same betting strategy and get the same cards for player and dealer you are drawing arbitrary lines in a continuous data stream. The difference is player A will bank win after win while my player B is very likely to bust his bankroll on his first session. Player A could adjust his betting strategy to misfortune while player B will over bet his bankroll as it is depleted increasing his RoR. If someone had a 100% RoR he should never hit his goal and either be play forever in one session or go broke, no other option. I dont see how one could view that as more profitable.

If player B played with a 60% RoR in your example of 10 unit BR to win 10 units as his goal.

Winning session total = .4 * (10) = 4
Losing session total = .6 * (-10 )= -6
Total sessions combined = -2

That looks like your higher RoR turned your +EV situation into a losing proposition to me. The math looks like elementary school level. In blackjack you lose more hands than you win. You rely on bonus blackjack payout and splitting and doubling to press your advantage(or lower your expected loss in some situations) to trim the HE down to a small number. Unless you only playing +EV situation(wonging) you play for a while at a disadvantage waiting for advantage to develop. If your bankroll is to small, you bust out before advantage can be developed cutting into your actual EV from the theoretical that you think you are playing with because this effect puts a negative bias to the average.

While increasing RoR can get a marginal bankroll to a more playable BR quicker it also makes it more likely that you will never get there. As you can see from my math it is a case of diminishing returns until the RoR finally takes your advantage away altogether.
 

Sharky

Well-Known Member
LovinItAll said:
...[I said] that lowering the ROR isn't necessarily the way to make the most money. ...
hey, nice work, this is REVOLUTIONARY!!!....quick, call the front desk and get me a table up here pronto...i want to play $100,000 units in my PJ's
 

tthree

Banned
Halving your goal of player A would have a smaller impact on RoR than doubling the goal of player B. Look at your example again and with player A having the same goal as player B (risk 10 to win 10) and make player B have a new goal of risk 10 to win 20. Maybe that will make it easier for you to visualize. Halving player A has a small impact on risk of ruin.
 

LovinItAll

Well-Known Member
tthree said:
Take your a example to an extreme and you will see the flaw in your logic. Make player B's goal be 10000. He will be playing many many days to win a session successfully and lose most sessions eventually if the RoR is high enough. Is player A going to stop gambling until B decides his session. I dont think by breaking it into sessions rather than time spent at the table is a legitimate comparison. If they use the same betting strategy and get the same cards for player and dealer you are drawing arbitrary lines in a continuous data stream. The difference is player A will bank win after win while my player B is very likely to bust his bankroll on his first session. Player A could adjust his betting strategy to misfortune while player B will over bet his bankroll as it is depleted increasing his RoR. If someone had a 100% RoR he should never hit his goal and either be play forever in one session or go broke, no other option. I dont see how one could view that as more profitable.

If player B played with a 60% RoR in your example of 10 unit BR to win 10 units as his goal.

Winning session total = .4 * (10) = 4
Losing session total = .6 * (-10 )= -6
Total sessions combined = -2

That looks like your higher RoR turned your +EV situation into a losing proposition to me. The math looks like elementary school level. In blackjack you lose more hands than you win. You rely on bonus blackjack payout and splitting and doubling to press your advantage(or lower your expected loss in some situations) to trim the HE down to a small number. Unless you only playing +EV situation(wonging) you play for a while at a disadvantage waiting for advantage to develop. If your bankroll is to small, you bust out before advantage can be developed cutting into your actual EV from the theoretical that you think you are playing with because this effect puts a negative bias to the average.

While increasing RoR can get a marginal bankroll to a more playable BR quicker it also makes it more likely that you will never get there. As you can see from my math it is a case of diminishing returns until the RoR finally takes your advantage away altogether.
There isn't a flaw in my logic. It's my problem, so I get to choose the parameters. ;). You said, "You rely on bonus blackjack payout and splitting and doubling to press your advantage:. I said, "In a game with an EV of +1%". I understand that this is a BJ board, but this wasn't necessarily a BJ question.

As I said before, I understand that RoR and profit have no correlation. I thought I would bop in here, ask a super simple question, get a bunch of "Of course that's right. Who doesn't get that?" answers, show my friend, and that would be that. Instead, I got two types of answers (one person actually 'got it' - I really appreciate that answer):

1. From smartasses who aren't nearly as smart as they think they are (life probably blows for them. It certainly does for those around them.).

2. Answers that were well-meaning, but changed the problem in such a way that the question wasn't answered at all.

If you were presented with this and had to make a decision:

"You are going to be given $1,000 each day for an indefinite period to bet on a BJ game that has an EV of +1% automagically built-in. Your bet for each hand must be $50. Standard blackjack rules apply (other than the 1% built-in advantage). You can either quit when you are up $200, $1000 (you must pick one at the beginning and you can't change when you bail after a win), or when you go broke. There are no other options."

Then you would have a choice to make: Settle for a smaller gain with more consistent results (smaller RoR), or make more money with wilder swings. One thing you wouldn't get to say is, "I reject the problem! Let's do it with $10,000 instead!" I've said it over and over - this question is a hypothetical, not a strategy. I was actually trying to prove exactly what everyone insists on trying to prove to me: that a lower RoR and increased 'profit' are not correlated. I happened to choose a data set that actually proved that the reverse can be true. I wasn't saying that I think it's a grand strategy.

@tthree: I know what you're trying to say, and I've understood that all along - it wasn't the problem, though. I do appreciate the answer. I think you're a good guy who actually wants to help other people.

We create arbitrary lines all of the time. Many people do not like to lose more than $x per day. They will quit playing no matter how good the game is. I know that no one will actually step up and agree with me, but there are MANY people here who have seen beautiful counts only to lose that day's "limit", even walking away while the shoe was favorable (maybe because they were broke, maybe because they just couldn't take it anymore - whatever). As a better example, we see it in poker all of the time, especially in the B&M rooms. A particular player's EV may be waaaaaay better than any advantage a BJ player will ever realize, yet he keeps getting sucked out on. Finally he just can't stand it anymore, starts to steam, and finally leaves, even though the game, over the longer term, is everything he could ask for. Doesn't mean he left broke, though, it just means that he couldn't stand the bad beats he was taking. Logic might say that he should keep playing because the game is so juicy, but he just can't do it.

Many people are trying to answer this from the perspective of a BJ player. It wasn't intended to be viewed through that narrow lens.

This wasn't a complex problem, it was a simple problem that apparently offended the sensibilities of some folks who couldn't/didn't want to get their head wrapped around the concept that there are questions where a decision must be made with the data that is given.

I'm done. I leave this thread knowing with 100% certainty that I'm correct re: the original problem, but respecting the well-meaning folks who misunderstood the problem and tried to take the time to show me my 'error' by changing the data set. To the people who jumped in with nothing useful to contribute (I think in the forum world they're called trolls), well, you have my sympathy, as you are exactly the kind of people that I try not to be. Life will always be worse for you than it needs to be. It's unfortunate, really.

Take care, and good luck at the tables. ~ LIA
 

tthree

Banned
On april 12, 2010 the wizard of odds posted the results of an experiment he did on a hypothetical game with a 1% (and 2%) player advantage that started with $100 and $1000 and the objective was to turn it into $1,000,000 using bet resizing as your bankroll fluctuated to bet 1% of your bankroll at all times. Here are the results:

$100, 1% player advantage
_________________________________________________________
Success (made $1,000,000 before busting): 79438 times (83.019%)
went bust: 16249 times (16.981%)
Average number of bets required to succeed: 174,972 (364.5 days @ 8 hrs/day 60 bets/hr)

===============================================

$1000, 1% player advantage
_____________________________________________________________
Success: 74,818 times (99.0285%)
went bust: 734 times (0.9715%)
Average number of bets: 137,208 (285.8 days @ 8 hrs/day 60 bets/hr)

Thats the raw data. Each time the $100 starting bankroll player added a significant digit (another zero. ie going from 100 to 1000) it would be about the same amount of time. So the time the risk of ruin was higher for the $100 starting bankroll would be from the start to $1000 mark (the first significant digit added) gradually approaching the same risk of ruin as the $1000 bankroll was 364.5/4 = 91.125 days. The $1000 bankroll added 3 significant digits in 285.8 days for a time duration of 95.267 days to grow your bankroll by one significant digit.

This shows the added risk of ruin indeed brought higher returns in the form of less time to reach a win goal. If the risk of ruin had remained constant rather than seeking the lesser risk of ruin gradually over time the difference would be much more pronounced. I hope that is what you were hoping to show your friend.

The other point bantered about in this thread the frequency of busting to grow the bankroll the same number of 4 significant digits was 16.981% for the $100 bankroll betting 1% of the bankroll at any time. The adjusted frequency of busting using the method above for the $1000 bankroll is 1.295% to grow it 4 significant digits betting 1% of the bankroll at any time. this is an approximation as this is not a linear function as it is in the above method.
 
Last edited:

LovinItAll

Well-Known Member
tthree said:
This shows the added risk of ruin indeed brought higher returns in the form of less time to reach a win goal. If the risk of ruin had remained constant rather than seeking the lesser risk of ruin gradually over time the difference would be much more pronounced. I hope that is what you were hoping to show your friend.
100% exactly my point, though I didn't state a time to reach a win goal. I was more than willing to concede that it may take longer to reach a stated win goal beyond the initial session goal, but that wasn't germaine.

Thank you. My position was to very simply - as I stated in my OP - show that RoR and profit were two entirely separate concepts and, more importantly, that lowering RoR could not possibly increase profits. I think that the idea of gradually adding to one's bankroll with very low RoR is sometimes confused with being "better" or "more profitable". It certainly will preserve one's bankroll, and for many, that's the most important component of their money management strategy. If everyone sought absolute minimum RoR, many would either be playing very small limits or could not actually find games small enough to play.

Take care ~ LIA
 
Top