Following the system's instructions, I place the ten pick-3 digits in a gird in order of most preferred digits to least. While it does have a 1 in 100 chance of hitting a straight and a 1 in 9 chance to match a front pair, it's designed for box hits. Those 10 combinations have a 1 in 9 chance of matching any drawing boxed when three different digits are drawn. For practical lottery play 100 drawings or $1000 bet should be sufficient.
I'm assuming you'll simulate 100 random bets of $10 in each of the 100 trials and to make the payoff structure easier, use the KY pick-3 payoff of $100 to $1. If some of random bets matches a double digit box hit in one of the probable 27 drawings that's a $200 to $1 payoff. Show us if this system can't beat chance with 6 hits or can't show a profit by with 11 hits.
A frequency digit system player would probably change the order at least once every 10 drawings and a player that keys on repeating digits might change the order every drawing. But to make the testing easier for you, use the same 10 combos in all 100 simulated drawing.
Looking forward to you giving us an actual example of what you started over a month ago!
4 5 6 had a box hit on today's KY midday drawing. Maybe I should look at that old system a little closer.
I'm not trying to get this pick-3 system back tested; just giving Jimmy a system so he show us the merits of back testing and simulating lottery systems.
mid-Ohio United States
Member #9
March 24, 2001
20,272 Posts
Offline
Quote: Originally posted by Stack47 on May 9, 2011
I'm not trying to get this pick-3 system back tested; just giving Jimmy a system so he show us the merits of back testing and simulating lottery systems.
Since the only rule is "place ten pick-3 digits in a gird in order of most preferred digits to least", the player is picking all the numbers and arranging them in any order he prefers. How is that a system that can be evaluated?
* you don't need to buy every combination, just the winning ones *
Kentucky United States
Member #32,651
February 14, 2006
10,302 Posts
Offline
Quote: Originally posted by RJOh on May 9, 2011
Since the only rule is "place ten pick-3 digits in a gird in order of most preferred digits to least", the player is picking all the numbers and arranging them in any order he prefers. How is that a system that can be evaluated?
I created the order using the frequency from the last ten combined KY midday and evening drawings. And to make things easier, Jimmy can use those ten combos in 100 trials. Most pick-3 systems change the combos weekly and some daily, but I just wanted to give Jimmy a chance with a real system to provide some valuable insights.
Following the system's instructions, I place the ten pick-3 digits in a gird in order of most preferred digits to least. While it does have a 1 in 100 chance of hitting a straight and a 1 in 9 chance to match a front pair, it's designed for box hits. Those 10 combinations have a 1 in 9 chance of matching any drawing boxed when three different digits are drawn. For practical lottery play 100 drawings or $1000 bet should be sufficient.
I'm assuming you'll simulate 100 random bets of $10 in each of the 100 trials and to make the payoff structure easier, use the KY pick-3 payoff of $100 to $1. If some of random bets matches a double digit box hit in one of the probable 27 drawings that's a $200 to $1 payoff. Show us if this system can't beat chance with 6 hits or can't show a profit by with 11 hits.
A frequency digit system player would probably change the order at least once every 10 drawings and a player that keys on repeating digits might change the order every drawing. But to make the testing easier for you, use the same 10 combos in all 100 simulated drawing.
Looking forward to you giving us an actual example of what you started over a month ago!
Stack47,
Your description of this old system raises more questions than it answers. If you refer to the code I posted in the Backtesting Thread recently, you will be in a much better position to program a simulation of it than me, since you seem to be familiar with it. Besides, you're asking the wrong questions.
When you make a request like, "Show us if this system can't beat chance with 6 hits or can't show a profit by with 11 hits.," it's not clear what you're asking for. "This system CAN show a profit year over year for 5 years" would be a more logical hypothesis, don't you think? "6 hits""11 hits" Of course, ANY system will appear successful in the short run IF it produces a specified number of wins! And what makes you think 100 drawings would be sufficient to provide confidence in the results? You really need to dig in to those 2 articles I linked to at the beginning of the Backtest and Simulation thread. It doesn't appear that you have, which makes me wonder if you're really serious about this.
Furthermore, and this is just one of several ambiguities in your description above, "...in order of most preferred digits to least." doesn't tell a programmer how you determine what makes one digit more preferred than another. Do you subscribe to the Gambler's Fallacy way of reckoning, or the Hot Hand school? If you expect anyone to have a chance at simulating your system correctly, you must first specify it more clearly, preferably using PseudoCode. A link to the original old system specification would also be helpful. If you did this, I'm sure you, or one of the other programmers here who believe they are far more skilled than me in this area, will be able to provide this programming service for you. I'll be glad to look over their code if they care to post it.
Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios? The answers are all there; I wish you could see them.
Here is an article that gives some very nice examples of how to express your algorithms in laymen's terms. The system challenges at The Wizard of Odds start with this sort of specification, so don't be insulted.
Kentucky United States
Member #32,651
February 14, 2006
10,302 Posts
Offline
Quote: Originally posted by jimmy4164 on May 10, 2011
Stack47,
Your description of this old system raises more questions than it answers. If you refer to the code I posted in the Backtesting Thread recently, you will be in a much better position to program a simulation of it than me, since you seem to be familiar with it. Besides, you're asking the wrong questions.
When you make a request like, "Show us if this system can't beat chance with 6 hits or can't show a profit by with 11 hits.," it's not clear what you're asking for. "This system CAN show a profit year over year for 5 years" would be a more logical hypothesis, don't you think? "6 hits""11 hits" Of course, ANY system will appear successful in the short run IF it produces a specified number of wins! And what makes you think 100 drawings would be sufficient to provide confidence in the results? You really need to dig in to those 2 articles I linked to at the beginning of the Backtest and Simulation thread. It doesn't appear that you have, which makes me wonder if you're really serious about this.
Furthermore, and this is just one of several ambiguities in your description above, "...in order of most preferred digits to least." doesn't tell a programmer how you determine what makes one digit more preferred than another. Do you subscribe to the Gambler's Fallacy way of reckoning, or the Hot Hand school? If you expect anyone to have a chance at simulating your system correctly, you must first specify it more clearly, preferably using PseudoCode. A link to the original old system specification would also be helpful. If you did this, I'm sure you, or one of the other programmers here who believe they are far more skilled than me in this area, will be able to provide this programming service for you. I'll be glad to look over their code if they care to post it.
Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios? The answers are all there; I wish you could see them.
Here is an article that gives some very nice examples of how to express your algorithms in laymen's terms. The system challenges at The Wizard of Odds start with this sort of specification, so don't be insulted.
P.S. Is this old system still in use or was it abandoned in the '80s?
"This system CAN show a profit year over year for 5 years"
That's very logical but only if the order of digits remained the same for 5 years and that's why I was being very liberal when I suggested 100 trials. Six hits in 100 drawings playing ten $1 box wagers I used beats chance and eleven hits would show a profit. This is a pencil and paper type system with lots of options so I chose one that could be tested. Since I used the combined midday and evening drawings digit frequency from the KY lottery, I can tell you the results of this group of ten box combos by July 1.
Any group of ten combos with 3 different digits has a 1 in 12 chance of hitting in the probable 72 drawings with 3 different digits and should average 6 hits in 100 trials. The ten combos using this system are mathematically arranged to always match 2 digits when 3 different digits are drawn. Since it will always match 2 digits, is it fallacy to believe this system has a 1 in 8 chance of hitting and should average 9 hits?
This group of combos may show many hits over 5 years but probably not enough to show a profit. To test for a consistent profit or just beating chance, a consistent method for choosing the digit order is necessary.
"Do you subscibe to the Gambler's Fallacy way of reckoning, or the Hot Hand school"
Gail Howard and others published books on how to wheel "your numbers" before PCs were common household items. Many of the articles and books on "average lottery play" written in the 80s can be found on the net when players didn't use PCs for data analysis. Their methods used the "hot hand" theory because they showed how the payoffs increased when "your numbers" were drawn.
Howard and others ideas probably evolved from players that used their favorite 12 numbers on two lines, matched all 6 numbers in a lotto game, but each line had 3 numbers and they won nothing. I thought about the pick-3 system when RJ showed that 12 RC5 numbers produced 4 jackpots over the history of that lottery.
"Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios?"
Nice work as usual but could have used some critical thinking on your test subjects. At least 5000 players will probably bet a $1 pick-3 QP somewhere tonight but will all them continue to bet $1 after 500 - 750 losses?
My knowledge of pick-3 play is based on working with many pick-3 players, speaking with people buying tickets, and what I've read on LP and other lottery based sites. I would be less critical of all of your work and opinions if they were based on more realistic play.
I found a book written the late 80s that does have examples of actual lottery play and discusses some of the system sellers.
The only casino gaming system I saw that could show huge profits was Edward O. Thorp's "Beat the Dealer" and that system only had a small window of opportunity before the casinos stopped end play, shuffled the cards more often, introduce multiple decks, and banned the erratic play suggested in his book. I've read many "how to play" roulette, craps, baccarat, and even some video poker and slot play books, but none of them claimed their methods would make the readers a fortune.
Basically to win the challenge a system must make at least $1 million in wagers on a starting $5000 bank roll and have at least $5001 when it stops after 200,000 trials. How many of your 5000 $1 QP players would make it to the last $1 bet after 20,000 drawings with a starting bankroll of $500?
I don't know if there are any winning systems and if there are some making claims, none of them will be taking that challenge.
"P.S. is this old system still in use or was it abandoned in the '80s?"
Since the this system is based on using 10 combos, each with three different digits, over time a player can only get a hit in 72% of the drawings. Continued play or successful play depends on the method used to enter the digits. Many players still use digit frequency so versions of it are being used today. What was unique about this system it use a dry erase board along with a template so it could be used over and over again.
NASHVILLE, TENN United States
Member #33,371
February 20, 2006
1,044 Posts
Offline
To address the question being asked, backtesting can be useful.
If one were to read of "+1-2+1" or the "Missing digists" system, then by backtesting you can determine if the idea has merit. If not, you move on.
To address the normal distribution curves at the beginning of this thread, there is no mathematical way to determine the numbers in the next draw. There is no magical equation. Paul Erdos has some unique ideas but I have yet to find anything which I can twist to fit the lottery. (Nor do I have an Erdos Number).
What we can do is take some of the concepts from Number Thoery and, with modification, apply the idea to the lottery. Now that would be a discussion in which I would love to participate.
United States
Member #93,943
July 10, 2010
2,180 Posts
Offline
Quote: Originally posted by Stack47 on May 10, 2011
"This system CAN show a profit year over year for 5 years"
That's very logical but only if the order of digits remained the same for 5 years and that's why I was being very liberal when I suggested 100 trials. Six hits in 100 drawings playing ten $1 box wagers I used beats chance and eleven hits would show a profit. This is a pencil and paper type system with lots of options so I chose one that could be tested. Since I used the combined midday and evening drawings digit frequency from the KY lottery, I can tell you the results of this group of ten box combos by July 1.
Any group of ten combos with 3 different digits has a 1 in 12 chance of hitting in the probable 72 drawings with 3 different digits and should average 6 hits in 100 trials. The ten combos using this system are mathematically arranged to always match 2 digits when 3 different digits are drawn. Since it will always match 2 digits, is it fallacy to believe this system has a 1 in 8 chance of hitting and should average 9 hits?
This group of combos may show many hits over 5 years but probably not enough to show a profit. To test for a consistent profit or just beating chance, a consistent method for choosing the digit order is necessary.
"Do you subscibe to the Gambler's Fallacy way of reckoning, or the Hot Hand school"
Gail Howard and others published books on how to wheel "your numbers" before PCs were common household items. Many of the articles and books on "average lottery play" written in the 80s can be found on the net when players didn't use PCs for data analysis. Their methods used the "hot hand" theory because they showed how the payoffs increased when "your numbers" were drawn.
Howard and others ideas probably evolved from players that used their favorite 12 numbers on two lines, matched all 6 numbers in a lotto game, but each line had 3 numbers and they won nothing. I thought about the pick-3 system when RJ showed that 12 RC5 numbers produced 4 jackpots over the history of that lottery.
"Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios?"
Nice work as usual but could have used some critical thinking on your test subjects. At least 5000 players will probably bet a $1 pick-3 QP somewhere tonight but will all them continue to bet $1 after 500 - 750 losses?
My knowledge of pick-3 play is based on working with many pick-3 players, speaking with people buying tickets, and what I've read on LP and other lottery based sites. I would be less critical of all of your work and opinions if they were based on more realistic play.
I found a book written the late 80s that does have examples of actual lottery play and discusses some of the system sellers.
The only casino gaming system I saw that could show huge profits was Edward O. Thorp's "Beat the Dealer" and that system only had a small window of opportunity before the casinos stopped end play, shuffled the cards more often, introduce multiple decks, and banned the erratic play suggested in his book. I've read many "how to play" roulette, craps, baccarat, and even some video poker and slot play books, but none of them claimed their methods would make the readers a fortune.
Basically to win the challenge a system must make at least $1 million in wagers on a starting $5000 bank roll and have at least $5001 when it stops after 200,000 trials. How many of your 5000 $1 QP players would make it to the last $1 bet after 20,000 drawings with a starting bankroll of $500?
I don't know if there are any winning systems and if there are some making claims, none of them will be taking that challenge.
"P.S. is this old system still in use or was it abandoned in the '80s?"
Since the this system is based on using 10 combos, each with three different digits, over time a player can only get a hit in 72% of the drawings. Continued play or successful play depends on the method used to enter the digits. Many players still use digit frequency so versions of it are being used today. What was unique about this system it use a dry erase board along with a template so it could be used over and over again.
'This system CAN show a profit year over year for 5 years'
You took this literally. It was only offered as an example of a hypothesis.
"That's very logical but only if the order of digits remained the same for 5 years and that's why I was being very liberal when I suggested 100 trials. Six hits in 100 drawings playing ten $1 box wagers I used beats chance and eleven hits would show a profit. This is a pencil and paper type system with lots of options so I chose one that could be tested. Since I used the combined midday and evening drawings digit frequency from the KY lottery, I can tell you the results of this group of ten box combos by July 1............."
You're completely missing my point. Perhaps it's my fault.
"At least 5000 players will probably bet a $1 pick-3 QP somewhere tonight but will all them continue to bet $1 after 500 - 750 losses?"
This exemplifies where you are going wrong in your thinking. This question is irrelevant. You're getting bogged down in details of a specific system. The important thing you need to see in the simulation output is the shape of the curve formed by the points surrounding the 50% Prize Ratio. 50,000 Quickpick players, which is essentially what the program is simulating, are perfectly capable of producing thousands of Prize Ratios less than 50% as well as thousands greater than 50% (beating chance) but only around 2% of them manage to come out ahead in terms of money with a Prize Ratio greater than 100% because of the low payouts.
'Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios?'
I think it would be helpful if you would make a simple change to the program and observe the shift in the frequency curve. Change the payout from $500 to $900 for the straight hit. Maybe then you will see what I'm driving at. Just change this line...
IF QP = LP THEN Equity = Equity + 500 ' to...
IF QP = LP THEN Equity = Equity + 900
The lotteries would still make 10%, as some of the online casinos do, and there would be a LOT more winners. However, overall, in the long run, the "average" person would lose 10%.
The bottom line is that a "system" is NOT necessary to produce the kind of variation of winning, losing, and patterns you are observing. I know, you think there exists a system that can put you consistently on the right hand side of that bell curve. I don't agree. From comments you've made here it doesn't appear you really have much confidence in the system you [sort of] described above, so how about proposing something that you believe has a chance at success?
United States
Member #93,943
July 10, 2010
2,180 Posts
Offline
Quote: Originally posted by GASMETERGUY on May 10, 2011
To address the question being asked, backtesting can be useful.
If one were to read of "+1-2+1" or the "Missing digists" system, then by backtesting you can determine if the idea has merit. If not, you move on.
To address the normal distribution curves at the beginning of this thread, there is no mathematical way to determine the numbers in the next draw. There is no magical equation. Paul Erdos has some unique ideas but I have yet to find anything which I can twist to fit the lottery. (Nor do I have an Erdos Number).
What we can do is take some of the concepts from Number Thoery and, with modification, apply the idea to the lottery. Now that would be a discussion in which I would love to participate.
I'm not ignoring you GASMETERGUY. I've just run out of time tonight. You're thinking in the right direction!
mid-Ohio United States
Member #9
March 24, 2001
20,272 Posts
Offline
After reading about the financial adviser winning the Hot Lotto jackpot by playing all 19 HB every drawing for three months I wanted to create a file to analyzes its data even though the game isn't played in Ohio.
I went to the PowerBall website to copy and paste the numbers but the page had a bunch of tabs at various places which made it difficult to convert the data to a file format I could use. I finally had to write a little program that looked at each character and delete the ones that weren't a number or space. The file had 948 drawings going back to 04/10/2002 so I wasn't looking forward to entering all that data by hand and the time to covert it to my file format was worth the extra effort.
I thought the total payout per dollar spent would be better than most games but it wasn't even though the overall odds of winning something was better than most. I may do a little more with the data later when I have the time.
* you don't need to buy every combination, just the winning ones *
Kentucky United States
Member #32,651
February 14, 2006
10,302 Posts
Offline
Quote: Originally posted by jimmy4164 on May 11, 2011
'This system CAN show a profit year over year for 5 years'
You took this literally. It was only offered as an example of a hypothesis.
"That's very logical but only if the order of digits remained the same for 5 years and that's why I was being very liberal when I suggested 100 trials. Six hits in 100 drawings playing ten $1 box wagers I used beats chance and eleven hits would show a profit. This is a pencil and paper type system with lots of options so I chose one that could be tested. Since I used the combined midday and evening drawings digit frequency from the KY lottery, I can tell you the results of this group of ten box combos by July 1............."
You're completely missing my point. Perhaps it's my fault.
"At least 5000 players will probably bet a $1 pick-3 QP somewhere tonight but will all them continue to bet $1 after 500 - 750 losses?"
This exemplifies where you are going wrong in your thinking. This question is irrelevant. You're getting bogged down in details of a specific system. The important thing you need to see in the simulation output is the shape of the curve formed by the points surrounding the 50% Prize Ratio. 50,000 Quickpick players, which is essentially what the program is simulating, are perfectly capable of producing thousands of Prize Ratios less than 50% as well as thousands greater than 50% (beating chance) but only around 2% of them manage to come out ahead in terms of money with a Prize Ratio greater than 100% because of the low payouts.
'Have you looked closely at my latest revision of the Pick-3 simulation in the other thread, the one that displays the results in terms of Prize Ratios?'
I think it would be helpful if you would make a simple change to the program and observe the shift in the frequency curve. Change the payout from $500 to $900 for the straight hit. Maybe then you will see what I'm driving at. Just change this line...
IF QP = LP THEN Equity = Equity + 500 ' to...
IF QP = LP THEN Equity = Equity + 900
The lotteries would still make 10%, as some of the online casinos do, and there would be a LOT more winners. However, overall, in the long run, the "average" person would lose 10%.
The bottom line is that a "system" is NOT necessary to produce the kind of variation of winning, losing, and patterns you are observing. I know, you think there exists a system that can put you consistently on the right hand side of that bell curve. I don't agree. From comments you've made here it doesn't appear you really have much confidence in the system you [sort of] described above, so how about proposing something that you believe has a chance at success?
Gotta run...
Jimmy4164
"It was only offered as an example of a hypothesis."
Five years is a very logical time period to test a system for profitably but this isn't a "play the same combos forever" type of system.
"This exemplifies where you are going wrong in your thinking."
I gave you the benefit of doubt by saying it's possible that 5000 players would purchase a $1 pick-3 QP tying to win $500 last night and I was being very liberal by saying someone would play the same ten combos for 100 drawings using that type of a system.
"You're getting bogged down in details of a specific system."
You must be assuming I believe risking $1000 with only an expected 72 chances to win in a 100 drawings to show a $100 profit is a good bet playing pick-3, but I don't. I only suggested testing that system because nobody else has provide a system for you to test.
My thinking is base on realistic lottery play and it's unrealistic for anyone to continue betting a $1 in every drawing using the same method of play trying to win $500 after losses over $500. Don't know the percentage of players having zero wins after 3 years of drawings but it's unrealistic to suggest even for test purposes the $1 a drawing bet trying to win $500 will continue for ten more years.
A more realistic approach is giving 5000 QP players a starting bankroll of $500 if $500 is the most they can expect to win on any one drawing.
"I know, you think there exists a system that can put you consistently on the right hand side of that bell curve. I don't agree."
Some states average over $1 million in pick-3 play every drawing with an average daily payoff of over $500,000. If there is no productive system, who is collecting the payoffs?
"From comments you've made here it doesn't appear you really have much confidence in the system you [sort of] described above,"
It was a simple pencil paper pick-3 system I saw over 20 years ago and thought it might be worth testing. A chart has ten 3 digit combo listed as "A through J". I checked the past 10 drawing digit frequency combining the KY midday and evening drawings and use the most frequent digit as letter "A" to the least frequent digit as letter "J".
I told you before I didn't like the idea of giving up 28% of the drawings by using three different digits in every drawing so that's why I never looked for possible playing potential.
"so how about proposing something that you believe has a chance at success?"
Hopefully someone will know which system that Iowa financial adviser used to win $9 million playing Hot Lotto and you test it.
United States
Member #93,943
July 10, 2010
2,180 Posts
Offline
Stack47,
What I hope I can do is help you to abandon your tendency to think about things like "people don't bet that way" or "nobody would play..." because if you don't, you will not be able to benefit from Monte Carlo (MC) simulations. There are many cases with MC where it's imperitive that the simulation eventually be modeled on real behaviour but for what we're doing here it doesn't matter at this point.
Before looking at short term results, please follow this thread: