Markov Chain theory
- Thread starter cdrake
- Start date
Hi CDrake and Martor,
I am interested in the application of Markov Chains to the lottery, but haven't come across anything that describes specifically how the theory should be applied.
According to Wikipedia:
So, this would seem to describe the random process of a lottery number as it changes from one state (e.g. 27) to another state (e.g. 32). Or am I off base here?
Anyway, I have created an excel spreadsheet that applies this process to lottery numbers by position.
For example, if the last draw had the number 7 in the first position, I can see from my spreadsheet that this has happened 94 times in 1,690 draws and that the transitions to the next draw are to a lower number 58.5% of the time, remaining the same (i.e. repeat in position) 5.3% of the time and transitioning to a higher number the remaining 36.2% of the time. I can also see that 43% of the time the number that followed 7 in the first position was between 1 and 4 and that 1 followed 7 19.4% of the time and 2 followed 7 10.8% of the time (as well a probability for every other number). I can see that no number higher than 22 has yet followed 7 in the first position.
But, even with all this information, the spreadsheet doesn't seem to be very helpful in predicting a specific number for the next draw, or even a tight range of numbers.
Is there something more that I should be doing here?
Martor -- you suggested combining with Bayes Theorem. I'm not a mathematician, so how would that work?
Looking forward to your insights,
I am interested in the application of Markov Chains to the lottery, but haven't come across anything that describes specifically how the theory should be applied.
According to Wikipedia:
So, this would seem to describe the random process of a lottery number as it changes from one state (e.g. 27) to another state (e.g. 32). Or am I off base here?
Anyway, I have created an excel spreadsheet that applies this process to lottery numbers by position.
For example, if the last draw had the number 7 in the first position, I can see from my spreadsheet that this has happened 94 times in 1,690 draws and that the transitions to the next draw are to a lower number 58.5% of the time, remaining the same (i.e. repeat in position) 5.3% of the time and transitioning to a higher number the remaining 36.2% of the time. I can also see that 43% of the time the number that followed 7 in the first position was between 1 and 4 and that 1 followed 7 19.4% of the time and 2 followed 7 10.8% of the time (as well a probability for every other number). I can see that no number higher than 22 has yet followed 7 in the first position.
But, even with all this information, the spreadsheet doesn't seem to be very helpful in predicting a specific number for the next draw, or even a tight range of numbers.
Is there something more that I should be doing here?
Martor -- you suggested combining with Bayes Theorem. I'm not a mathematician, so how would that work?
Looking forward to your insights,
Hi Icewynd. I just new from some of the trials that I had conducted with the Markov utility, that I downloaded from Ions site years ago, that most of the combinations produced from the program were heavily weighted towards those numbers in the most recent draws. Sometimes I would get 3 or 4 number hits and a lot of times I wouldn't get any. It all comes down to consistency for me.
cdrake said:
So you are saying that Saliu's program seems to drop numbers that haven't come out recently? That makes sense.
I'm not sure if his stuff is still available, but I hate having to go into DOS to run his programs, so probably won't bother. I guess I've gotten spoiled by all the bells and whistles produced by modern software.
Good luck!