Practical lessons from a craps simulator....
November 12th, 2013 at 9:23:02 AM
permalink
Probabilities, standard deviations, variance, and expectations are nice terms and fun for the academic...but their practical application for making money on the craps table is very, very limited. Having finished my excel simulator, I've spent a lot of time using it to answer different questions. None of the questions included different betting strategies. Rather, I have been more interested in validating what I felt to be true empirically, ie., from experience. Again, this was not about various betting strategies. For simplicity, presume the pass line or don't pass line flat bet. Also, I will speak in fairly general terms, so if I say it's about 30% and you feel the need to point out that it's 29.6%...don't bother. The following observations are not in any significant order.
First...dismiss the idea that you'll be at a table long enough to complete even one 495 come out roll session. It takes about 1600 +/- 100 rolls to generate 494 pass line decisions (the first roll of the day is a freebie). At a generous 100 rolls per hour, that's 15-17 hours of play. On a weekend at a crowded table where 50 rolls per hour is generous...30-34 hours.
Second...never expect your results to mirror the rule of 495 in any way except that at some point you'll have a good chance of being down seven bets. Maybe I should say, don't count on your results to mirror...you can expect all you want, but don't count on or budget your money based on the "expected" results. The people that quote the million roll sample size to prove the 495 results....well, practically speaking...give it a rest. From the paragraph above that means 10,000-20,000 hours of play. Good luck with that.
Third...a huge sample size is misleading to the real time bettor. As above, over a gazillion rolls the difference between wins and losses is seven bets. Mathematically speaking, the difference is seven bets. Employ the previous two paragraphs to realize that in reality, it will be coincidence or choice if you walk away from your table down seven bets.
Fourth...small practical sample sizes provide more realistic views about what the results might be. That's right, small sample sizes. If you're stuck leaning on those probabilities and expected outcomes, and too paralyzed by them to make a bet...run the simulator. Run it for 36 rolls...run it for 36 rolls, and another 36, and another 36. You'll be extremely lucky if you ever see 36 consecutive rolls produce numbers exactly as their probabilities say they should show. However, the individual numbers do tend to stay within 2 standard deviations of their expected appearances. In other words, don't be worried about the 12 showing up 10 times in a row. Can it happen, sure! Is it likely to...that's your bet.
Fifth...there are plenty of opportunities to walk away from the table a winner. The key is walking away while you're up. Yes, over a gazillion rolls you'll be a small loser, but during those gazillion rolls you'll be everywhere between a big loser and a big winner. The question is how many rolls (how much time) are you going to play? When you run a number of smaller sample sizes, run them based on the amount of rolls per hour you think is average for your tables. This will give you an excellent idea of how different each session can be.
Nothing I've said is news to the real-time player. As a bettor, knowing the gazillion roll outcome means nothing. If you're a newer player and want to see how different length sessions chart out, make a simulator. Some statisticians think they know the game because they know the probabilities, expected outcomes and everything in between. They don't. The game is about what's happening right now. Keep your simulator session lengths similar to what you're likely to experience. Make a mental note of how different the outcomes are. Sometimes the 7/11 come out wins are plentiful, sometimes they're nearly absent. For example, don't make the mistake of thinking that 30% of rolls are come out rolls so out of 50 rolls there will be 15 come out rolls and 22% of those will be 7/11 winners so almost 4 of those rolls will be pass line winners.
In the end what you'll see is as I mentioned above, there are plenty of opportunities to make money at craps and walk away a winner. The only part of it that isn't luck (variance) is when you decide to walk away. When (if) you run a simulator, you'll see that it's luck (variance) if you're betting the pass line when the come out 7/11's hit. It's luck (variance) if your points hit. It's luck (variance) if you place the 6/8 and hit a few times before the 7 shows. It's luck (variance) if you wait for a 5 count and make a come bet and the 7/11 shows. The one and only real hope you should have is that you're betting during a time when the dice roll in your favor. All the stats in the world will not help you win at the table, they won't help you walk away a winner. The simulator won't help you win more or walk away a winner either. A simulator will help you realize that, at the table, statistical probabilities and expectations mean nothing. Luck (variance) makes money but won't tell you when to walk away. That's up to you.
Edit: The word "variance" has been added in parenthesis where I used the word 'luck' as a result of gr8player's very nice practical application of variance.
First...dismiss the idea that you'll be at a table long enough to complete even one 495 come out roll session. It takes about 1600 +/- 100 rolls to generate 494 pass line decisions (the first roll of the day is a freebie). At a generous 100 rolls per hour, that's 15-17 hours of play. On a weekend at a crowded table where 50 rolls per hour is generous...30-34 hours.
Second...never expect your results to mirror the rule of 495 in any way except that at some point you'll have a good chance of being down seven bets. Maybe I should say, don't count on your results to mirror...you can expect all you want, but don't count on or budget your money based on the "expected" results. The people that quote the million roll sample size to prove the 495 results....well, practically speaking...give it a rest. From the paragraph above that means 10,000-20,000 hours of play. Good luck with that.
Third...a huge sample size is misleading to the real time bettor. As above, over a gazillion rolls the difference between wins and losses is seven bets. Mathematically speaking, the difference is seven bets. Employ the previous two paragraphs to realize that in reality, it will be coincidence or choice if you walk away from your table down seven bets.
Fourth...small practical sample sizes provide more realistic views about what the results might be. That's right, small sample sizes. If you're stuck leaning on those probabilities and expected outcomes, and too paralyzed by them to make a bet...run the simulator. Run it for 36 rolls...run it for 36 rolls, and another 36, and another 36. You'll be extremely lucky if you ever see 36 consecutive rolls produce numbers exactly as their probabilities say they should show. However, the individual numbers do tend to stay within 2 standard deviations of their expected appearances. In other words, don't be worried about the 12 showing up 10 times in a row. Can it happen, sure! Is it likely to...that's your bet.
Fifth...there are plenty of opportunities to walk away from the table a winner. The key is walking away while you're up. Yes, over a gazillion rolls you'll be a small loser, but during those gazillion rolls you'll be everywhere between a big loser and a big winner. The question is how many rolls (how much time) are you going to play? When you run a number of smaller sample sizes, run them based on the amount of rolls per hour you think is average for your tables. This will give you an excellent idea of how different each session can be.
Nothing I've said is news to the real-time player. As a bettor, knowing the gazillion roll outcome means nothing. If you're a newer player and want to see how different length sessions chart out, make a simulator. Some statisticians think they know the game because they know the probabilities, expected outcomes and everything in between. They don't. The game is about what's happening right now. Keep your simulator session lengths similar to what you're likely to experience. Make a mental note of how different the outcomes are. Sometimes the 7/11 come out wins are plentiful, sometimes they're nearly absent. For example, don't make the mistake of thinking that 30% of rolls are come out rolls so out of 50 rolls there will be 15 come out rolls and 22% of those will be 7/11 winners so almost 4 of those rolls will be pass line winners.
In the end what you'll see is as I mentioned above, there are plenty of opportunities to make money at craps and walk away a winner. The only part of it that isn't luck (variance) is when you decide to walk away. When (if) you run a simulator, you'll see that it's luck (variance) if you're betting the pass line when the come out 7/11's hit. It's luck (variance) if your points hit. It's luck (variance) if you place the 6/8 and hit a few times before the 7 shows. It's luck (variance) if you wait for a 5 count and make a come bet and the 7/11 shows. The one and only real hope you should have is that you're betting during a time when the dice roll in your favor. All the stats in the world will not help you win at the table, they won't help you walk away a winner. The simulator won't help you win more or walk away a winner either. A simulator will help you realize that, at the table, statistical probabilities and expectations mean nothing. Luck (variance) makes money but won't tell you when to walk away. That's up to you.
Edit: The word "variance" has been added in parenthesis where I used the word 'luck' as a result of gr8player's very nice practical application of variance.
...in the long run, you'll lose money.
November 12th, 2013 at 10:17:41 AM
permalink
Hello, nowakezone. Great post.
I should mention that a.) I rarely post outside of my own threads and b.) I am not a craps player; but, all of that aside, I felt compelled to respond the intelligence and practicality that permeates your post.
Knowing when you've hit your variance, be it of the high or the low variety, and when to walk afterwards, is of paramount importance to the serious player.
Want to replace the word "luck" for the term "variance"? Fine. Matters not to me.
Either way, one must be able to determine their best time to walk a winner, or, for that matter, a loss as well.
It is, as nowakezone so profoundly put forth, about all that we can control and, hence, maximize to our advantage.
Once again, great post, my friend.
I should mention that a.) I rarely post outside of my own threads and b.) I am not a craps player; but, all of that aside, I felt compelled to respond the intelligence and practicality that permeates your post.
Knowing when you've hit your variance, be it of the high or the low variety, and when to walk afterwards, is of paramount importance to the serious player.
Want to replace the word "luck" for the term "variance"? Fine. Matters not to me.
Either way, one must be able to determine their best time to walk a winner, or, for that matter, a loss as well.
It is, as nowakezone so profoundly put forth, about all that we can control and, hence, maximize to our advantage.
Once again, great post, my friend.
November 12th, 2013 at 10:37:24 AM
permalink
Quote: gr8playerHello, nowakezone. Great post.
I should mention that a.) I rarely post outside of my own threads and b.) I am not a craps player; but, all of that aside, I felt compelled to respond the intelligence and practicality that permeates your post.
Knowing when you've hit your variance, be it of the high or the low variety, and when to walk afterwards, is of paramount importance to the serious player.
Want to replace the word "luck" for the term "variance"? Fine. Matters not to me.
Either way, one must be able to determine their best time to walk a winner, or, for that matter, a loss as well.
It is, as nowakezone so profoundly put forth, about all that we can control and, hence, maximize to our advantage.
Once again, great post, my friend.
It comes down to one word "CONTROL". Yes, control yourself. Know when to quit(up or down)
November 12th, 2013 at 1:02:37 PM
permalink
Very kind of you gr8player. Hey I really like your practical application of variance. I hadn't connected those dots between variance and luck, so thank you for making the connection for me. As such, I'm going to edit my post a bit to reflect your help. Thanks again.
...in the long run, you'll lose money.
November 12th, 2013 at 1:24:22 PM
permalink
Quote: nowakezoneVery kind of you gr8player. Hey I really like your practical application of variance. I hadn't connected those dots between variance and luck, so thank you for making the connection for me. As such, I'm going to edit my post a bit to reflect your help. Thanks again.
I too liked your post, well said.
Determining when to walk I like to refer to as discipline.
