Kentucky United States Member #32652 February 14, 2006 7308 Posts Offline

Posted: December 19, 2012, 4:55 pm - IP Logged

Quote: Originally posted by Ronnie316 on December 19, 2012

Ok, so Boney will be supportive if I change the thread name to..................

"Do some number combinations get better results"

I thought that's what you meant when you started this thread over six months ago. We agreed 98,280 QPs have the same odds as a group of 28 numbers and distinguished how the odds changed only when both matched five numbers.

On page 5 you said "The objective is to at least TRY to hit to a JP." and I have no idea why the discussion drifted from that goal.

New Jersey United States Member #99032 October 18, 2010 1439 Posts Offline

Posted: December 19, 2012, 4:56 pm - IP Logged

Quote: Originally posted by Stack47 on December 19, 2012

I already did.

"We could calculate the probability of the number of times each MM number should be drawn over the next 700 drawings and see it's around 59 to 65 times, but when looking at the last 700 drawings we'll see only 30% of the numbers matched that probability."

The mean is 59 to 65 but we'll round if off to 62. What is the standard deviation?

Oh I missed that.

Keep in mind it's been a while since I've had to use this formula so I may be missing a step... I doubt it though. (Most of the work has already been done by others in my gambling so I have little need of a formula to find STD DEV for something like the average Lottery numbers, I basically use the kelly criterion, if anyone was interested)

I also don't remember the actual proper test for this.... it's more complicated than the one you use if negatives are possible. I don't really want to go through and do a lot of work for this, so I'm going to use very rough estimates rather than take a couple hours to go through the formulas relevant to this type of test....

Well the mean is 700/(1/(5/56)), which simplifies to 62.5. The variance is approximately 1/(5/56) or 11.2. (I use an approximate b/c doing the math with every possible outcome is tedious) The square root of that is STD DEV, so the STD DEV for one trial is around 3.34. The STD DEV for multiple trials is the square root of the number of trials times STD DEV for one trial. So Squareroot of 700 times 3.34 = about 88.3. Since negative values are meaningless (none can be less than 0) the expectation for any idividual number is to come up between 18.35-106.65 - 68.2% of the time.

Looking back at my work I know I made an error, most likely because I think I was supposed use the Chi squared test, which honestly, I completely forgot how to do. I realize you're going to ridicule me for that, but I don't really care b/c I know it's out there and if you wanted to get an accurate answer, someobody could dig it up. Basically the reason for this is that a number can't come up less than 0 times, so the whole test should really be Right Skewed (if you looked at the results on a curve) creating a situation where the STD DEV measurement has a different definition than with a curve that isn't skewed.

So it's probably more accurate to say that there would likely be more STD DEV would be smaller if I performed the proper test. Anyway, that's my long, and partly innacurate answer. Although it's safe to say that not abornmal for numbers to be pretty far off of there expected average until there have been many draws. Say what you will about the differences in numbers' frequencies from their expected average, I say it's completely explainable through the proper statistics. Unless it was consistently way off the average, I wouldn't be concerned.

New Jersey United States Member #99032 October 18, 2010 1439 Posts Offline

Posted: December 19, 2012, 4:57 pm - IP Logged

Quote: Originally posted by Stack47 on December 19, 2012

I thought that's what you meant when you started this thread over six months ago. We agreed 98,280 QPs have the same odds as a group of 28 numbers and distinguished how the odds changed only when both matched five numbers.

On page 5 you said "The objective is to at least TRY to hit to a JP." and I have no idea why the discussion drifted from that goal.

New Jersey United States Member #99032 October 18, 2010 1439 Posts Offline

Posted: December 19, 2012, 4:59 pm - IP Logged

Quote: Originally posted by Ronnie316 on December 19, 2012

Ok, so Boney will be supportive if I change the thread name to..................

"Do some number combinations get better results"

Just answering your question in the title.

But I will stop now, since I've made my point many times. That point being that no number combinations have better odds.

If I was going to answer that question, then I'd have to say that yes, some combination get better results. In some regards because you can look back and see who won in the past, and because you can use those results to slightly decrease your chances of sharing a jackpot, and in parimutual states, increase all the other prizes as well.

But I don't believe the odds change in any meaningful and detectable way.

If anybody quotes me and asks me to reply I will, otherwise I won't other than to check on the results of our challenge.

United States Member #116268 September 7, 2011 20244 Posts Offline

Posted: December 19, 2012, 5:29 pm - IP Logged

Quote: Originally posted by Boney526 on December 19, 2012

Just answering your question in the title.

But I will stop now, since I've made my point many times. That point being that no number combinations have better odds.

If I was going to answer that question, then I'd have to say that yes, some combination get better results. In some regards because you can look back and see who won in the past, and because you can use those results to slightly decrease your chances of sharing a jackpot, and in parimutual states, increase all the other prizes as well.

But I don't believe the odds change in any meaningful and detectable way.

If anybody quotes me and asks me to reply I will, otherwise I won't other than to check on the results of our challenge.

You are too funny Boney.......

You can post here all you want. You make a lot of good points, but the stated lottery odds will always be "overall odds" and not actually odds. Some people just have a way of getteing BETTER ODDS.........

United States Member #116268 September 7, 2011 20244 Posts Offline

Posted: December 19, 2012, 5:35 pm - IP Logged

Quote: Originally posted by Boney526 on December 19, 2012

Fair enough, onwards with your goal, then!

Once again you either fail to see the point, or refuse to acknowledge it............

A jackpot has ALREADY been hit in the challenge we did for 39 draws......

People like LottoBoner dont have any problem SEEING that because I was looking for something to happen and predicting accordingly the 5+1 match was made.............

Not because I effected the outcome, but because I anticipated it.........

Kentucky United States Member #32652 February 14, 2006 7308 Posts Offline

Posted: December 19, 2012, 6:18 pm - IP Logged

Quote: Originally posted by Boney526 on December 19, 2012

Oh I missed that.

Keep in mind it's been a while since I've had to use this formula so I may be missing a step... I doubt it though. (Most of the work has already been done by others in my gambling so I have little need of a formula to find STD DEV for something like the average Lottery numbers, I basically use the kelly criterion, if anyone was interested)

I also don't remember the actual proper test for this.... it's more complicated than the one you use if negatives are possible. I don't really want to go through and do a lot of work for this, so I'm going to use very rough estimates rather than take a couple hours to go through the formulas relevant to this type of test....

Well the mean is 700/(1/(5/56)), which simplifies to 62.5. The variance is approximately 1/(5/56) or 11.2. (I use an approximate b/c doing the math with every possible outcome is tedious) The square root of that is STD DEV, so the STD DEV for one trial is around 3.34. The STD DEV for multiple trials is the square root of the number of trials times STD DEV for one trial. So Squareroot of 700 times 3.34 = about 88.3. Since negative values are meaningless (none can be less than 0) the expectation for any idividual number is to come up between 18.35-106.65 - 68.2% of the time.

Looking back at my work I know I made an error, most likely because I think I was supposed use the Chi squared test, which honestly, I completely forgot how to do. I realize you're going to ridicule me for that, but I don't really care b/c I know it's out there and if you wanted to get an accurate answer, someobody could dig it up. Basically the reason for this is that a number can't come up less than 0 times, so the whole test should really be Right Skewed (if you looked at the results on a curve) creating a situation where the STD DEV measurement has a different definition than with a curve that isn't skewed.

So it's probably more accurate to say that there would likely be more STD DEV would be smaller if I performed the proper test. Anyway, that's my long, and partly innacurate answer. Although it's safe to say that not abornmal for numbers to be pretty far off of there expected average until there have been many draws. Say what you will about the differences in numbers' frequencies from their expected average, I say it's completely explainable through the proper statistics. Unless it was consistently way off the average, I wouldn't be concerned.

Just so you know, the 766,584 groups that match five numbers in 5 consecutive drawings in a 39 drawing period only represent a tiny percentage of the total number of 28 number groups. The total is a 46 digit number.

The standard deviation for density (variance) is 11 and only 5 numbers are outside the parameters in 700 drawings.

United States Member #116268 September 7, 2011 20244 Posts Offline

Posted: December 19, 2012, 6:36 pm - IP Logged

Quote: Originally posted by Stack47 on December 19, 2012

Just so you know, the 766,584 groups that match five numbers in 5 consecutive drawings in a 39 drawing period only represent a tiny percentage of the total number of 28 number groups. The total is a 46 digit number.

The standard deviation for density (variance) is 11 and only 5 numbers are outside the parameters in 700 drawings.

Ok, heres my "quirky" set of 39 numbers for tonights PB drawing. Wed. Dec. 18, 2012.

United States Member #124493 March 14, 2012 7023 Posts Offline

Posted: December 19, 2012, 6:43 pm - IP Logged

Quote: Originally posted by Stack47 on December 19, 2012

I thought that's what you meant when you started this thread over six months ago. We agreed 98,280 QPs have the same odds as a group of 28 numbers and distinguished how the odds changed only when both matched five numbers.

On page 5 you said "The objective is to at least TRY to hit to a JP." and I have no idea why the discussion drifted from that goal.

United States Member #124493 March 14, 2012 7023 Posts Offline

Posted: December 19, 2012, 6:44 pm - IP Logged

Quote: Originally posted by Stack47 on December 19, 2012

I already did.

"We could calculate the probability of the number of times each MM number should be drawn over the next 700 drawings and see it's around 59 to 65 times, but when looking at the last 700 drawings we'll see only 30% of the numbers matched that probability."

The mean is 59 to 65 but we'll round if off to 62. What is the standard deviation?

United States Member #124493 March 14, 2012 7023 Posts Offline

Posted: December 19, 2012, 6:53 pm - IP Logged

Quote: Originally posted by LottoBoner on December 19, 2012

So do the calculation boney. Whats the standard deviation of the results.

You are seeing it all wrong. Instead of one game you have to think of it as a bunch of little games. Should I bet on the number that is out 59 games, the number out 1 game or the number at 32 games.

What is the STD DEV of the Mean, The arithemetic mean, the mode, what is the percentages?

Shoud I bet on the number with 62 hits in a year or the one with 30 hits.

United States Member #124493 March 14, 2012 7023 Posts Offline

Posted: December 19, 2012, 7:49 pm - IP Logged

Quote: Originally posted by LottoBoner on December 19, 2012

So do the calculation boney. Whats the standard deviation of the results.

You are seeing it all wrong. Instead of one game you have to think of it as a bunch of little games. Should I bet on the number that is out 59 games, the number out 1 game or the number at 32 games.

What is the STD DEV of the Mean, The arithemetic mean, the mode, what is the percentages?

Shoud I bet on the number with 62 hits in a year or the one with 30 hits.

What is the STD DEV GAP JAP GRO MOO MIN MAX

What is the expectancy?

I also don't remember the actual proper test for this.... it's more complicated than the one you use if negatives are possible. I don't really want to go through and do a lot of work for this, so I'm going to use very rough estimates rather than take a couple hours to go through the formulas relevant to this type of test....

Well the mean is 700/(1/(5/56)), which simplifies to 62.5. The variance is approximately 1/(5/56) or 11.2. (I use an approximate b/c doing the math with every possible outcome is tedious) The square root of that is STD DEV, so the STD DEV for one trial is around 3.34. The STD DEV for multiple trials is the square root of the number of trials times STD DEV for one trial. So Squareroot of 700 times 3.34 = about 88.3. Since negative values are meaningless (none can be less than 0) the expectation for any idividual number is to come up between 18.35-106.65 - 68.2% of the time.

Looking back at my work I know I made an error, most likely because I think I was supposed use the Chi squared test, which honestly, I completely forgot how to do. I realize you're going to ridicule me for that, but I don't really care b/c I know it's out there and if you wanted to get an accurate answer, someobody could dig it up. Basically the reason for this is that a number can't come up less than 0 times, so the whole test should really be Right Skewed (if you looked at the results on a curve) creating a situation where the STD DEV measurement has a different definition than with a curve that isn't skewed.

So it's probably more accurate to say that there would likely be more STD DEV would be smaller if I performed the proper test. Anyway, that's my long, and partly innacurate answer. Although it's safe to say that not abornmal for numbers to be pretty far off of there expected average until there have been many draws. Say what you will about the differences in numbers' frequencies from their expected average, I say it's completely explainable through the proper statistics. Unless it was consistently way off the average, I wouldn't be concerned

New Jersey United States Member #99032 October 18, 2010 1439 Posts Offline

Posted: December 20, 2012, 12:24 am - IP Logged

Quote: Originally posted by LottoBoner on December 19, 2012

I also don't remember the actual proper test for this.... it's more complicated than the one you use if negatives are possible. I don't really want to go through and do a lot of work for this, so I'm going to use very rough estimates rather than take a couple hours to go through the formulas relevant to this type of test....

Well the mean is 700/(1/(5/56)), which simplifies to 62.5. The variance is approximately 1/(5/56) or 11.2. (I use an approximate b/c doing the math with every possible outcome is tedious) The square root of that is STD DEV, so the STD DEV for one trial is around 3.34. The STD DEV for multiple trials is the square root of the number of trials times STD DEV for one trial. So Squareroot of 700 times 3.34 = about 88.3. Since negative values are meaningless (none can be less than 0) the expectation for any idividual number is to come up between 18.35-106.65 - 68.2% of the time.

Looking back at my work I know I made an error, most likely because I think I was supposed use the Chi squared test, which honestly, I completely forgot how to do. I realize you're going to ridicule me for that, but I don't really care b/c I know it's out there and if you wanted to get an accurate answer, someobody could dig it up. Basically the reason for this is that a number can't come up less than 0 times, so the whole test should really be Right Skewed (if you looked at the results on a curve) creating a situation where the STD DEV measurement has a different definition than with a curve that isn't skewed.

So it's probably more accurate to say that there would likely be more STD DEV would be smaller if I performed the proper test. Anyway, that's my long, and partly innacurate answer. Although it's safe to say that not abornmal for numbers to be pretty far off of there expected average until there have been many draws. Say what you will about the differences in numbers' frequencies from their expected average, I say it's completely explainable through the proper statistics. Unless it was consistently way off the average, I wouldn't be concerned

Like I said I'm not going to go study the Chi Squared distribution test just to find the STD DEV, Mode, etc. of that.

Take it as you will. I'm not gonna go and study Statistics for a couple hours just to remember and do the Chi Squared test only for you to respond with emoticons. Clearly you can see how that's a huge waste of my time.

United States Member #93947 July 10, 2010 2180 Posts Offline

Posted: December 20, 2012, 12:53 am - IP Logged

Quote: Originally posted by Boney526 on December 19, 2012

Oh I missed that.

Keep in mind it's been a while since I've had to use this formula so I may be missing a step... I doubt it though. (Most of the work has already been done by others in my gambling so I have little need of a formula to find STD DEV for something like the average Lottery numbers, I basically use the kelly criterion, if anyone was interested)

I also don't remember the actual proper test for this.... it's more complicated than the one you use if negatives are possible. I don't really want to go through and do a lot of work for this, so I'm going to use very rough estimates rather than take a couple hours to go through the formulas relevant to this type of test....

Well the mean is 700/(1/(5/56)), which simplifies to 62.5. The variance is approximately 1/(5/56) or 11.2. (I use an approximate b/c doing the math with every possible outcome is tedious) The square root of that is STD DEV, so the STD DEV for one trial is around 3.34. The STD DEV for multiple trials is the square root of the number of trials times STD DEV for one trial. So Squareroot of 700 times 3.34 = about 88.3. Since negative values are meaningless (none can be less than 0) the expectation for any idividual number is to come up between 18.35-106.65 - 68.2% of the time.

Looking back at my work I know I made an error, most likely because I think I was supposed use the Chi squared test, which honestly, I completely forgot how to do. I realize you're going to ridicule me for that, but I don't really care b/c I know it's out there and if you wanted to get an accurate answer, someobody could dig it up. Basically the reason for this is that a number can't come up less than 0 times, so the whole test should really be Right Skewed (if you looked at the results on a curve) creating a situation where the STD DEV measurement has a different definition than with a curve that isn't skewed.

So it's probably more accurate to say that there would likely be more STD DEV would be smaller if I performed the proper test. Anyway, that's my long, and partly innacurate answer. Although it's safe to say that not abornmal for numbers to be pretty far off of there expected average until there have been many draws. Say what you will about the differences in numbers' frequencies from their expected average, I say it's completely explainable through the proper statistics. Unless it was consistently way off the average, I wouldn't be concerned.

"...I basically use the kelly criterion, if anyone was interested..."

I tried to introduce this a long time ago - you can see the enthusiasm it generated here...