- Home
- Premium Memberships
- Lottery Results
- Forums
- Predictions
- Lottery Post Videos
- News
- Search Drawings
- Search Lottery Post
- Lottery Systems
- Lottery Charts
- Lottery Wheels
- Worldwide Jackpots
- Quick Picks
- On This Day in History
- Blogs
- Online Games
- Premium Features
- Contact Us
- Whitelist Lottery Post
- Rules
- Lottery Book Store
- Lottery Post Gift Shop
The time is now 8:25 am
You last visited
June 3, 2024, 8:15 am
All times shown are
Eastern Time (GMT-5:00)
Bit PickingPrev TopicNext Topic
-
I'm not much of a gambler, but for whatever reason the NY Pick 10 (Keno) game caught my attention way back in 2001. I was convinced that I could crack the code and set about attempting to do so by doing the computer equivalent of counting on my fingers as I am no mathematician, but had access to plenty of computing power and storage.
16 years later I still have not cracked the code, but I am way ahead versus the amount of money I have put in. I've spent most of the last 16 years living outside of NY so I have not been playing the game and been theorizing as a hobby. Although truth be told even if I had been living in NY it is doubtful that I would have continued to play on a daily basis because the idea of just picking numbers at random and hoping that they hit offends my senses.
I forget exactly when I discovered this forum, but I did spend a significant amount of time browsing the forums for ideas. I am a private person so it is with some hesitation that I'm posting some of the things that I have learned over the years, but I do believe in giving back which is why I have decided to write this post.
The reason why P10 appealed to me was because every day there were 184, 756 winning combinations that appeared each day and I was convinced that within that number there must be a pattern that I could find. Finding that pattern has been far more challenging than I thought it would be.
With P10 the average number appears at an interval of roughly once every four days over an extended period of time. On average 4 numbers that appeared the previous day will appear again the next day. On average two to three numbers that appeared 4 days prior will appear again.
Seeing these patterns most of my effort focused on identifying potential numbers based on three basic statistics
Hits Per Period (HPP): Since numbers appeared roughly every for days I would use periods that were always divisible by 4. For example I might use 4, 12, 28, 60 and 100 to segment the numbers
Separation (SEP): This is the number of days since the last time the number appeared as a winner.
In a Row (IAR): This is the number of times the number has appeared consecutively in
I even extended the statistics to include the last 10 times a number had appeared with stats of lets say a SEP of 4, and IAR of 0 and HPP100 (100 days back) of 24 and then hit or missed the next day. Most percentages came within a point or two of the 25% average and while I could see variations that might push the likelihood of a number hitting up to 60% or as low as a fraction of a percent, the winning numbers were never the number with the highest possible probability of hitting. They were always some mixture of high, medium and low. At best I was able to narrow it down to 1 about 1 million. While that is better than 1 in 8,911,711 it was still relatively unrealistic odds.
The more I continued to experiment the more I came to dislike the notion of picking by number. There were just too many variables as well as a certain amount of human bias regarding certain number sequences so I began focusing on methods that provided me with little to no insight as to what the combination of numbers was going to be.
With pick 10 there are 1,646,492,110,120 combinations of 10 digit numbers. What I decided to do was to convert the number values to their binary equivalent. To accommodate all the possible variations would require 41 different bits (2 ^ 41). For example the numbers 1,7,14,33,55,60,61,72, 75,80 look like this in binary : 00001011111110110101011011100100100000110.
Personally I was much more comfortable picking 1 and 0 than having to choose numbers so I instantly took a liking to it and most of my methods have focused around this. Plus this method accommodates a wide of games. For instance the average pick 5 out of 30 game only requires 18 bits.
However it really shines with P10 because due to all of the patterns that show up each day a majority of the winning numbers have exactly 20 1's. of The range can vary from as low as 10 all the way up through about 33, but 20 is quite consistent and offers a wide range of strategies. Because of the number of results that had 20 1's I could pick a result that supported the pattern that I wanted to establish.
For my purposes I ignored the first bit or sometimes committed to playing to that value set to 0 or 1 so that I was always working with an even number. Also because 2^41 is a much larger number range than the actual number of combinations some of the combinations of bits did not translate into valid combinations when the first bit was set to 1.
This methodology opened a up wide range of strategies. My favorite was to force a pattern where there were 20 bits that had an IAR of 1, 10 that had an IAR of 2, 5 that had an IAR of 3 , 3 that had an IAR of 4 and 2 that had an IAR of 5 with the same pattern in place for the 0's. If I looked at the results over the next 7 days there was always almost at least one pattern that would appear where half the bits from each section would graduate to the next IAR level while the other half would be flipped to the opposite value.
I spent a lot of time trying to tweak this strategy, but the percentage of times I actually won using this strategy when I would run long term scenarios did not make it worth the risk of playing any real money. The problem with these kinds of strategies is that while they provide a chance to pick the actual winner you have t be prepared to play for possibly a number of years before hitting that jackpot and during that time the return on the money being spent is about 1/3 to 1/2 depending on how many time you hit 9 out of 10.
Another method that I explored was what i refer to as Bit IAR. Basically every number is made of of blocks where the are 1,2,3,4, etc. bits in a row where the value is set to one. One of the most common patterns was 3 blocks of 1, 2 blocks of 2, 1 block of 3 and 1 block of 4. What I liked about this method was that by time I got to placing the blocks of 3 and 4 there were only a limited number of positions where they could be placed.
Below are some screen shots of how the selector worked. Depending on which bits I selected the numbers I was going to wind up in playing would change accordingly.
Now that I am back in NY I have not decided on what strategy I will use and experimenting with some additional ones, but at some point I will decide on one and run with it for a little while because winning this thing is on my bucket list.
Hopefully you will find this posting helpful in your own endeavors.
-
Philadelpia/Pennsylvania
United States
Member #2,218
September 1, 2003
6,805 Posts
OfflineQuote: Originally posted by Novan60 on Feb 24, 2017
I'm not much of a gambler, but for whatever reason the NY Pick 10 (Keno) game caught my attention way back in 2001. I was convinced that I could crack the code and set about attempting to do so by doing the computer equivalent of counting on my fingers as I am no mathematician, but had access to plenty of computing power and storage.
16 years later I still have not cracked the code, but I am way ahead versus the amount of money I have put in. I've spent most of the last 16 years living outside of NY so I have not been playing the game and been theorizing as a hobby. Although truth be told even if I had been living in NY it is doubtful that I would have continued to play on a daily basis because the idea of just picking numbers at random and hoping that they hit offends my senses.
I forget exactly when I discovered this forum, but I did spend a significant amount of time browsing the forums for ideas. I am a private person so it is with some hesitation that I'm posting some of the things that I have learned over the years, but I do believe in giving back which is why I have decided to write this post.
The reason why P10 appealed to me was because every day there were 184, 756 winning combinations that appeared each day and I was convinced that within that number there must be a pattern that I could find. Finding that pattern has been far more challenging than I thought it would be.
With P10 the average number appears at an interval of roughly once every four days over an extended period of time. On average 4 numbers that appeared the previous day will appear again the next day. On average two to three numbers that appeared 4 days prior will appear again.
Seeing these patterns most of my effort focused on identifying potential numbers based on three basic statistics
Hits Per Period (HPP): Since numbers appeared roughly every for days I would use periods that were always divisible by 4. For example I might use 4, 12, 28, 60 and 100 to segment the numbers
Separation (SEP): This is the number of days since the last time the number appeared as a winner.
In a Row (IAR): This is the number of times the number has appeared consecutively in
I even extended the statistics to include the last 10 times a number had appeared with stats of lets say a SEP of 4, and IAR of 0 and HPP100 (100 days back) of 24 and then hit or missed the next day. Most percentages came within a point or two of the 25% average and while I could see variations that might push the likelihood of a number hitting up to 60% or as low as a fraction of a percent, the winning numbers were never the number with the highest possible probability of hitting. They were always some mixture of high, medium and low. At best I was able to narrow it down to 1 about 1 million. While that is better than 1 in 8,911,711 it was still relatively unrealistic odds.
The more I continued to experiment the more I came to dislike the notion of picking by number. There were just too many variables as well as a certain amount of human bias regarding certain number sequences so I began focusing on methods that provided me with little to no insight as to what the combination of numbers was going to be.
With pick 10 there are 1,646,492,110,120 combinations of 10 digit numbers. What I decided to do was to convert the number values to their binary equivalent. To accommodate all the possible variations would require 41 different bits (2 ^ 41). For example the numbers 1,7,14,33,55,60,61,72, 75,80 look like this in binary : 00001011111110110101011011100100100000110.
Personally I was much more comfortable picking 1 and 0 than having to choose numbers so I instantly took a liking to it and most of my methods have focused around this. Plus this method accommodates a wide of games. For instance the average pick 5 out of 30 game only requires 18 bits.
However it really shines with P10 because due to all of the patterns that show up each day a majority of the winning numbers have exactly 20 1's. of The range can vary from as low as 10 all the way up through about 33, but 20 is quite consistent and offers a wide range of strategies. Because of the number of results that had 20 1's I could pick a result that supported the pattern that I wanted to establish.
For my purposes I ignored the first bit or sometimes committed to playing to that value set to 0 or 1 so that I was always working with an even number. Also because 2^41 is a much larger number range than the actual number of combinations some of the combinations of bits did not translate into valid combinations when the first bit was set to 1.
This methodology opened a up wide range of strategies. My favorite was to force a pattern where there were 20 bits that had an IAR of 1, 10 that had an IAR of 2, 5 that had an IAR of 3 , 3 that had an IAR of 4 and 2 that had an IAR of 5 with the same pattern in place for the 0's. If I looked at the results over the next 7 days there was always almost at least one pattern that would appear where half the bits from each section would graduate to the next IAR level while the other half would be flipped to the opposite value.
I spent a lot of time trying to tweak this strategy, but the percentage of times I actually won using this strategy when I would run long term scenarios did not make it worth the risk of playing any real money. The problem with these kinds of strategies is that while they provide a chance to pick the actual winner you have t be prepared to play for possibly a number of years before hitting that jackpot and during that time the return on the money being spent is about 1/3 to 1/2 depending on how many time you hit 9 out of 10.
Another method that I explored was what i refer to as Bit IAR. Basically every number is made of of blocks where the are 1,2,3,4, etc. bits in a row where the value is set to one. One of the most common patterns was 3 blocks of 1, 2 blocks of 2, 1 block of 3 and 1 block of 4. What I liked about this method was that by time I got to placing the blocks of 3 and 4 there were only a limited number of positions where they could be placed.
Below are some screen shots of how the selector worked. Depending on which bits I selected the numbers I was going to wind up in playing would change accordingly.
Now that I am back in NY I have not decided on what strategy I will use and experimenting with some additional ones, but at some point I will decide on one and run with it for a little while because winning this thing is on my bucket list.
Hopefully you will find this posting helpful in your own endeavors.
Thanks for your information.
What program are you using?
-
I think you might be my long lost twin, Welcome to LP and love your post.
RL
....
-
Welcome to LP!
Gambler667
-
Quote: Originally posted by winsumloosesum on Feb 24, 2017
Thanks for your information.
What program are you using?
I originally started with an Access Database as that was all that I had available to me for storing, querying the data and generating play slips. Eventually I graduated to a combination of SQL Server and VB programming in Visual Studio which offers a lot more flexibility.
-
Quote: Originally posted by Novan60 on Feb 24, 2017
I'm not much of a gambler, but for whatever reason the NY Pick 10 (Keno) game caught my attention way back in 2001. I was convinced that I could crack the code and set about attempting to do so by doing the computer equivalent of counting on my fingers as I am no mathematician, but had access to plenty of computing power and storage.
16 years later I still have not cracked the code, but I am way ahead versus the amount of money I have put in. I've spent most of the last 16 years living outside of NY so I have not been playing the game and been theorizing as a hobby. Although truth be told even if I had been living in NY it is doubtful that I would have continued to play on a daily basis because the idea of just picking numbers at random and hoping that they hit offends my senses.
I forget exactly when I discovered this forum, but I did spend a significant amount of time browsing the forums for ideas. I am a private person so it is with some hesitation that I'm posting some of the things that I have learned over the years, but I do believe in giving back which is why I have decided to write this post.
The reason why P10 appealed to me was because every day there were 184, 756 winning combinations that appeared each day and I was convinced that within that number there must be a pattern that I could find. Finding that pattern has been far more challenging than I thought it would be.
With P10 the average number appears at an interval of roughly once every four days over an extended period of time. On average 4 numbers that appeared the previous day will appear again the next day. On average two to three numbers that appeared 4 days prior will appear again.
Seeing these patterns most of my effort focused on identifying potential numbers based on three basic statistics
Hits Per Period (HPP): Since numbers appeared roughly every for days I would use periods that were always divisible by 4. For example I might use 4, 12, 28, 60 and 100 to segment the numbers
Separation (SEP): This is the number of days since the last time the number appeared as a winner.
In a Row (IAR): This is the number of times the number has appeared consecutively in
I even extended the statistics to include the last 10 times a number had appeared with stats of lets say a SEP of 4, and IAR of 0 and HPP100 (100 days back) of 24 and then hit or missed the next day. Most percentages came within a point or two of the 25% average and while I could see variations that might push the likelihood of a number hitting up to 60% or as low as a fraction of a percent, the winning numbers were never the number with the highest possible probability of hitting. They were always some mixture of high, medium and low. At best I was able to narrow it down to 1 about 1 million. While that is better than 1 in 8,911,711 it was still relatively unrealistic odds.
The more I continued to experiment the more I came to dislike the notion of picking by number. There were just too many variables as well as a certain amount of human bias regarding certain number sequences so I began focusing on methods that provided me with little to no insight as to what the combination of numbers was going to be.
With pick 10 there are 1,646,492,110,120 combinations of 10 digit numbers. What I decided to do was to convert the number values to their binary equivalent. To accommodate all the possible variations would require 41 different bits (2 ^ 41). For example the numbers 1,7,14,33,55,60,61,72, 75,80 look like this in binary : 00001011111110110101011011100100100000110.
Personally I was much more comfortable picking 1 and 0 than having to choose numbers so I instantly took a liking to it and most of my methods have focused around this. Plus this method accommodates a wide of games. For instance the average pick 5 out of 30 game only requires 18 bits.
However it really shines with P10 because due to all of the patterns that show up each day a majority of the winning numbers have exactly 20 1's. of The range can vary from as low as 10 all the way up through about 33, but 20 is quite consistent and offers a wide range of strategies. Because of the number of results that had 20 1's I could pick a result that supported the pattern that I wanted to establish.
For my purposes I ignored the first bit or sometimes committed to playing to that value set to 0 or 1 so that I was always working with an even number. Also because 2^41 is a much larger number range than the actual number of combinations some of the combinations of bits did not translate into valid combinations when the first bit was set to 1.
This methodology opened a up wide range of strategies. My favorite was to force a pattern where there were 20 bits that had an IAR of 1, 10 that had an IAR of 2, 5 that had an IAR of 3 , 3 that had an IAR of 4 and 2 that had an IAR of 5 with the same pattern in place for the 0's. If I looked at the results over the next 7 days there was always almost at least one pattern that would appear where half the bits from each section would graduate to the next IAR level while the other half would be flipped to the opposite value.
I spent a lot of time trying to tweak this strategy, but the percentage of times I actually won using this strategy when I would run long term scenarios did not make it worth the risk of playing any real money. The problem with these kinds of strategies is that while they provide a chance to pick the actual winner you have t be prepared to play for possibly a number of years before hitting that jackpot and during that time the return on the money being spent is about 1/3 to 1/2 depending on how many time you hit 9 out of 10.
Another method that I explored was what i refer to as Bit IAR. Basically every number is made of of blocks where the are 1,2,3,4, etc. bits in a row where the value is set to one. One of the most common patterns was 3 blocks of 1, 2 blocks of 2, 1 block of 3 and 1 block of 4. What I liked about this method was that by time I got to placing the blocks of 3 and 4 there were only a limited number of positions where they could be placed.
Below are some screen shots of how the selector worked. Depending on which bits I selected the numbers I was going to wind up in playing would change accordingly.
Now that I am back in NY I have not decided on what strategy I will use and experimenting with some additional ones, but at some point I will decide on one and run with it for a little while because winning this thing is on my bucket list.
Hopefully you will find this posting helpful in your own endeavors.
Seeing these patterns most of my effort focused on identifying potential numbers based on three basic statistics
How bias can one be in locating these patterns? and how do connect the above statement with the next:
The more I continued to experiment the more I came to dislike the notion of picking by number. There were just too many variables as well as a certain amount of human bias regarding certain number sequences so I began focusing on methods that provided me with little to no insight as to what the combination of numbers was going to be.
Personally, seeing is purely observational, so most patterns are located after the fact.The second statement is more in tune with the game's premise, ie, you Generate unobservable patterns than trying to locate bias patterns.
-
Quote: Originally posted by adobea78 on Feb 26, 2017
Seeing these patterns most of my effort focused on identifying potential numbers based on three basic statistics
How bias can one be in locating these patterns? and how do connect the above statement with the next:
The more I continued to experiment the more I came to dislike the notion of picking by number. There were just too many variables as well as a certain amount of human bias regarding certain number sequences so I began focusing on methods that provided me with little to no insight as to what the combination of numbers was going to be.
Personally, seeing is purely observational, so most patterns are located after the fact.The second statement is more in tune with the game's premise, ie, you Generate unobservable patterns than trying to locate bias patterns.
The bias is not in the locating the patterns. The bias comes into play when it comes time to select the numbers as almost every strategy is about limiting choice to the smallest set of numbers with the highest probability of appearing together.
For example if a strategy produces a number like 1,2,3,4,28 the average person will be hesitant to play that number because the odds of four sequential numbers appearing together in the same drawing offends our senses. So if I was using my strategy to pick a number for the 4th position and my high probability choices were 4,5,8 and 13 I would instinctively avoid choosing the number 4 when that very well may be the number that is drawn.
What I liked about choosing by bits is that every position is set to on or off. What numbers were going to appear based on my bit selection had no bearing based on which bits were set to on or off. And because every drawing had a result where 20 bits were set to 1 I could focus my energies on patterns centered around the odds of how often a bit would be consecutively be on or off.
Here is another way to think of picking by bits. If I take the total number of combinations and divide them by two each day (depending on if the total number of combinations was evenly divisible by 2) I would have 50% chance of guessing whether or not the winning number would come from the upper or lower half of the combination tree.
Once I make a decision on the first half I can then split the next section in half and keep repeating the process until I have one digit left. NY P6 required 26 bits. In my testing using a 5 year long playing strategy at $10.00 a drawing I was consistently able to get at least 15 of the 26 bits that I needed at least once over the length of the strategy time frame. In theory by playing the other 2048 combinations could have guaranteed a win within that 5 year time frame.
The problem is that even with a win in adding that extra money to each drawing I could wind up in spending more to win than what I actually won if I won on a day when the jackpot reset or had to split it with someone else. That made it a non-viable strategy.
Some would argue that I would have been better off just spending the 20,000 some odd dollars per drawing it would have taken to play that strategy on quick picks. The problem with that is that it would only have reduced my odds to about 1 in 1000 with no guarantee for a win after 1000 drawing were completed. Even though the other strategy was not viable it at least guaranteed a jackpot payout the 1000 times I tested it.
Granted on the 1001 time it could have failed, but 1000 cycles is usually enough to confirm a trend.
-
Might I suggest you take a step back and consider an alternative approach.
Choose one drawing.
Track data for this one drawing only.
For example, I play Mega Millions only on Friday. So, I only track Friday drawings.
The Tuesday drawing is ignored.
Basic outline:
Divide the game into 8 Decades.
A thru H
R for integer 80
etc
Create an Alpha Signature for each 10 integers per Decade per drawing.
4.9.13.14.27.28.35.39.48.50 = AABBCCDDEF
You can track these and will no doubt discover a few patterns giving suggestions for future plays.
etc
Create Game History charts for each Decade.
Divide each Decade into three Main groups.Decade A
Main A - 123
Main B - 456
Main C - 789
Decade B
Main A - 10 11 12
Main B - 13 14 15
Main C - 16-17-18-19
etc
There is more to it, of course, but I don't think I have to dot every I and cross every T.
Tracking will eventually show where the current emphasis is, given that the game is moving forward from draw to draw.
When choosing 10 integers to play, maybe you would pick at least 1 integer from each Main, depending on game history.
Maybe the trend lines will suggest 3 integers from same Main would be good choice.
Over a period of time you will have a good idea of where the game for the drawing you choose has been and where it might be headed.
No algorithms, no mathematical formulae, no odds,no probability, no percentage, no fancy statistical charts, no odds, no evens, no highs, no lows,no roots, no heads, no tails, no pairs, no triads, no vtrac.
Just a good set of trend data, a functional brain and ability to interpret trends and make good choices. -
Quote: Originally posted by bobby623 on Feb 26, 2017
Might I suggest you take a step back and consider an alternative approach.
Choose one drawing.
Track data for this one drawing only.
For example, I play Mega Millions only on Friday. So, I only track Friday drawings.
The Tuesday drawing is ignored.
Basic outline:
Divide the game into 8 Decades.
A thru H
R for integer 80
etc
Create an Alpha Signature for each 10 integers per Decade per drawing.
4.9.13.14.27.28.35.39.48.50 = AABBCCDDEF
You can track these and will no doubt discover a few patterns giving suggestions for future plays.
etc
Create Game History charts for each Decade.
Divide each Decade into three Main groups.Decade A
Main A - 123
Main B - 456
Main C - 789
Decade B
Main A - 10 11 12
Main B - 13 14 15
Main C - 16-17-18-19
etc
There is more to it, of course, but I don't think I have to dot every I and cross every T.
Tracking will eventually show where the current emphasis is, given that the game is moving forward from draw to draw.
When choosing 10 integers to play, maybe you would pick at least 1 integer from each Main, depending on game history.
Maybe the trend lines will suggest 3 integers from same Main would be good choice.
Over a period of time you will have a good idea of where the game for the drawing you choose has been and where it might be headed.
No algorithms, no mathematical formulae, no odds,no probability, no percentage, no fancy statistical charts, no odds, no evens, no highs, no lows,no roots, no heads, no tails, no pairs, no triads, no vtrac.
Just a good set of trend data, a functional brain and ability to interpret trends and make good choices.This is a nice idea, but in general I have tried to move away from any strategies that involve picking 1 out of X numbers. While I was quite enthusiastic about picking by bits I did eventually move on from it because trying to accurately guess 41 yes/no values proved too challenging. The reason why i posted it is because it might spark an idea in someone else that takes them down a completely different path.
One of the reasons why I started using what I refer to as the BIT IAR method was because it reduced the number of yes/no choices that I had to make, but I abandoned it because it had morphed back into a picking a 1 in X exercise.
However going through that process gave me some other ideas and my efforts are primarily focused on finding ways of making yes/no decisions in as few bits as possible to get a playable number. The less choices I have to make the less likely I ma to make the wrong choice.
The trade off for reducing choices means I have to be willing to accept a longer interval between the frequency of the pattern. The trick is finding a pattern that strikes the right balance between minimizing choice, the number of games required to play each drawing and the amount of time it takes for the pattern to cycle.
-
Novan60
Did you ever try cutting down your bits then use a method similar to the way resistor color coding works. Using multiplier bits
for millions, thousands hundreds etc... and a division bit. I have worked on several different methods but have yet to put any
of them into a completed software package. I am not familiar with the BIT IAR so maybe you have considered this already.
The things you mentioned above would be nice but in the end you can't shorten the odds. It's possible to work out the best
numbers to play, just don't expect them to show in the next game. Sometimes chance will favor us and sometimes not. While
we can pick up a few numbers with good analysis chance always has the final say. Enjoy the game and the risk.
RL
....
-
Quote: Originally posted by RL-RANDOMLOGIC on Feb 27, 2017
Novan60
Did you ever try cutting down your bits then use a method similar to the way resistor color coding works. Using multiplier bits
for millions, thousands hundreds etc... and a division bit. I have worked on several different methods but have yet to put any
of them into a completed software package. I am not familiar with the BIT IAR so maybe you have considered this already.
The things you mentioned above would be nice but in the end you can't shorten the odds. It's possible to work out the best
numbers to play, just don't expect them to show in the next game. Sometimes chance will favor us and sometimes not. While
we can pick up a few numbers with good analysis chance always has the final say. Enjoy the game and the risk.
RL
BIT IAR is a name that I made up to define what I was trying to accomplish with this strategy. Bit Blocks would probably be a more appropriate name but I had already used that name for a different strategy. I try to be descriptive in my strategy names so that when I look back to an old strategy I can recall what I was trying to accomplish without having to work my way through the code or write extensive notes.
Converting a set of numbers to bits was my first exploration into a new strategy in this branch. Being able to translate a number into bits wound up in opening the door to about two dozen or so different strategies that I have tried with most of them revolving around ways to figure out how to get a winning result from having to make as few yes/no choices as possible.
BIT IAR was just one way that I explored but abandoned because it took me away from making yes/no choices and had turned it into a pick 7 out of 41 with the additional requirement that blocks had to be in certain positions.
Most of the rest of the strategies that i tried revolved around establishing a baseline and just guessing where the changes are going to take place. For instance if I eliminate the first bit, divide the remaining 40 by 4 and compare that to a baseline number from a previous drawing there will be at least one result where each segment had to positions that changed bits.
This is possible because in picking 2 out of 4 digits there are 6 possibilities. 6^10 = 60,466,176 different combinations of numbers. The odds of picking a winning number in the game ar 1 in 8,911,711 which means that in by having over 60 million combinations I have covered the minimum odds 7 times over.
This method can be further honed in by numbering in bit in each segment from 1 to 4 and deciding that every winning change result is going to have X number of 1's, 2s, 3's and 4's. For example I might decide that a winning pattern is going to have 5 ones, 5 twos, 5 threes and 5 fours. From there it is now just a matter of picking which segments get which numbers and I once I have used up all of one number I know that I can't choose it anymore.
So far the best I've been able to do is get the number of choices down to 20 yes/no choices which gives me about a million different combinations of numbers to work with and in the pick 10 game that is a large enough number set to ensure a winner every X number of days.
20 is a lot better than 41, but the consequence of narrowing down the choices is that it directly impacts the frequency in which the pattern occurs. The only way to counter that is to rearrange the base data to conform to a trend. For instance instead of looking at the bits sequentially, reorder them based on the frequency in which they have appeared as 1 over the last X days in the baseline set.
In many ways moving away from NY has saved me a ton of cash because instead of rushing to develop a new strategy and running out to play it I wound up in spending much more time building data models and running through long term outcomes using all of the historical data that I have collected. Now that I am back in NY I will eventually start playing again, but am in no rush as I have some ideas that I still want to work through before committing real money again.
-
I've mentioned picking by bits as the foundation for developing a strategy to pick a winning set of numbers, here is a walk through on developing the foundation of a strategy for picking numbers this way. For this example I will use a P5B30 game. In this game there 142,506 number combinations. This means it will require 18 bits in order to generate all of the possible number combinations.
It is important to keep in mind that 2^18 = 262,144 so that means that there are some combinations of bits that will exceed the number of valid possibilities so we have to be mindful whenever the first bit is set to 1. This is why I usually just play both values for that bit.
Mathematics will tell us that 9 bits set to 1 appears in about 19% of the possible numbers combinations which means that statistically speaking we should expect to see a winning number set with 9 bets once every 5 days.
I always like to see how the actual math translates into real drawings. The following is a report on how many bits appeared in each drawing and it would appear that the 9 bits every 5 or so drawings holds true. We also see that number sets involving 0 to 4 bits or 13 and higher do not appear often enough to make them worth considering as viable number choices.
One of the reasons I try to sample as large a data set as possible is that a small sampling can skew the results. Here are the results if I just look at a 50 day window:
50 seems like a fairly large number to sample with, but it still allows for hot/cold numbers to skew the data which could result in the development of a faulty strategy. Had I just relied on this sample I would have picked 7 as my target and while it is not a bad option it is not the best one and would impact strategy choices later on.
Because statistics basically reset every drawing I like to look at patterns of change. This is one of the reasons I like bits over of methodologies based on hot and cold numbers.
I like to start from a baseline value that represents the most likely number of bits so that I am working with a consistent baseline value. Math and the data indicate that value should be 9.
For this strategy I will be looking at drawings where the number of bits was 9. Since it is impossible to predict what numbers are going to appear in the next drawing I'm going to look at the next X drawing from my baseline drawing. This game allows me to submit a game slip for the next 26 drawings so that is the range I will use for the development of my strategy.
At this point I will construct a table that has all of my baseline values.
When I look at the baseline table I see that the interval at which 9 bits appear varies, but eventually evens out to about once every 5 days. So from a budget standpoint I should plan on spending 26 dollars every 5 days or so.
The other benefit of building this table is that I can use the data to try and identify possible trends.
Since my strategy is dependent on picking numbers based on the bits that have changed from my baseline number I will want to decide on a target number of bits to change. Once again math tells us that most likely number of bits to change will be 9. That information is also validated in the actual historical data.
Based on math and what the historical data has proved I could play a strategy where I target number combinations where 9 bits are different from my baseline bits and that for every 26 drawings that I play I have a 99% chance that a winning set will show and that on average I will have four opportunities to win per 26 drawings. This means that my raw odds are 1 in 48,620.
Technically the odds are about 1 in 35,000 or so because not every change combination of 9 bits works out to be a viable number depending on what the base bits are. While those odds still do not guarantee a win, they are certainly better than the base odds and I'm willing to pass up the possibility of winning a majority of the drawing if I can triple my likelihood of guessing correctly on certain days.
One could argue that I could take that same 26 dollars and get my odds down to 1 in 5,481 for a single drawing which is not bad odds and certainly better than 1 in 35,000 although considering that on average there are 4 changes to win per 26 drawings it really is 1 in 9,000 and a strategy that is much more sustainable over a longer period of time.
If I did play 26 dollars a day on quick picks hopefully at some point I might win. However if I make the same 26 dollar a day commitment to this strategy the odds become 1 in 1,346 on the days that match the pattern in the strategy and would be 1 in 51 for the entire cycle although my costs would then jump to 676 dollars per cycle.
Keep in mind that even with the odds being 1 in 51 it is entirely possible I could go through 51 cycles, spend $34,476.00 and not win the jackpot. In most cases that exceeds the average jackpot before taxes for these kinds of games so it means that it is not worth the risk. When looking at strategies it is important to also factor in worst case scenarios because nothing is guaranteed in games of chance.
I could take this strategy a step further by specifying how many 1 bits from the baseline get changed to 0 and how many 0 bits get changed to 1. Math tells us that it will be a 4/5 split and running through the historical data proves this to be true.
So what is that we gain by doing this? Since our baseline number has 9 bits set to 1 if we choose the 4/5 option this means that our odds of picking the correct pattern are 1 in 126 for each side lowering our odds down to 1 in 15,876. In actuality they are more likely to be about 1 in 11,000 by time invalid combinations are removed.
There was however a price to pay for the improved odds, we have gone from a 99% chance each cycle to an 84% chance and the average occurrence has been reduced from 4 times a cycle to just 2. To me it is worth the 15 percentage points if it means that I am picking from a smaller set of possibilities because all strategies are still going to be dependent on a certain amount of luck.
There is more that I could do to refine this strategy to eventually get the odds down to a base of about 1 in 1000 with a 50% chance of appearing within each cycle. The important part is that I can see how each choice I make impacts my possibilities of winning while gauging how much and how long I would have to spend in order to properly execute the strategy to a winning result.
Hopefully you have found this helpful and can incorporate some of these ideas into your own strategy planning.