Welcome Guest
You last visited January 18, 2017, 7:05 pm
All times shown are
Eastern Time (GMT-5:00)

# Probability of a MegaMillions rollover.

Topic closed. 352 replies. Last post 10 years ago by Prob988.

 Page 3 of 24
Greenwich, CT
United States
Member #4793
May 24, 2004
1822 Posts
Offline
 Posted: November 5, 2005, 5:12 pm - IP Logged

Yes, thank you.

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 7, 2005, 12:22 pm - IP Logged

Mostly just to keep my mathematical muscles working, I did some more difficult work on exponential curve fitting and, playing around with the model, I came up with a function that matches the behavior of lottery pot growth pretty well, at least in the region in which we now are.  The mathematical details of this approach can be found on the internet, http://mathworld.wolfram.com/LeastSquaresFittingExponential.html.  As you can see the model is fairly complex computationally, but I had some other modelling spreadsheets available that I could modify appropriately.

To get this function to work well, I had to use only three data points to generate the curve fit, and I used data from the last large run, and not this run.  Still, I think it is superior to my last model.  The function that I was able to generate is: S = \$561,000*exp(0.317*N) where S is the sales for the particular drawing and N is the number of the drawing, including the first drawing, which is N=1.  The exp is the exponential function.  (This function does not fit sales behavior in the early drawings very well, but those incorporating more data points, that fit early behavior better, do not model behavior of late drawings very well.)

Using this model I have constructed the following long range probabilities for rollovers from this point on.  Here is the table for what I came up with:

Annuity Value                  Cash Value                    Single draw              Prob of all

Rollover Prob.          Rollovers

 \$733,420,458.18 \$427,013,689 16.64% 0.36% \$561,321,837.59 \$326,814,048 27.07% 2.18% \$435,918,446.74 \$253,801,407 38.59% 8.05% \$344,540,547.36 \$200,599,163 49.97% 20.85% \$277,956,059.98 \$161,832,195 60.32% 41.73% \$229,437,829.75 \$133,583,803 69.19% 69.19%

The last column is the probability that all the jackpot in question and all of the previous jackpots, including the one we have going on right now (note that the computed values are very close to the advertised values) will rollover.  Thus we have a 40% chance of seeing a jack pot with a \$200M cash value, and a 20% chance of seeing a jackpot with an annuity value of over \$400M.  The chances of larger jackpots are small, but not negligible.

I think it's a better model and it will be interesting to see if it works well, should we have more rollovers.

BILOXI, MISSISSIPPI
United States
Member #19651
August 3, 2005
621 Posts
Offline
 Posted: November 7, 2005, 12:42 pm - IP Logged

looks good the roll overs seem right but still no luck for a 150 million roll

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 7, 2005, 12:51 pm - IP Logged

looks good the roll overs seem right but still no luck for a 150 million roll

My model is just a model.  Somehow I think that if a jackpot over \$300M rolled - or maybe even before - my model would breakdown or at least become less accurate.  Nothing about what I have written should be taken as absolute.  It just gives one an idea of what the situation is approximately.

That said, I think it would take an extraordinary run to jump the jackpot by \$150M.  The record jump - I think - and Jake is better equipped than I to tell this, is the jump from \$230 to \$363 million in early May 2000 for the old big game.  Players have become much more jaded since then.

Member #2673
November 2, 2003
497 Posts
Offline
 Posted: November 7, 2005, 1:41 pm - IP Logged

Yes, the record jump is as you stated Prob988 - \$133 million in May 2000.

Unfortunately we cannot know for certain if that jump was entirely based on sales or was it partially based on the poor estimate of the \$230 million jackpot. I have noticed that Mega Millions (and Big Game) never adjusted their estimates after the draw to the actual value based upon sales. Therefore the \$230 million jackpot may have grown to say \$263 million by the time of the draw. Under this case, the jump was only \$100 million.

Powerball on the other hand always states the actual jackpot value after the draw even if the jackpot was not won. A questinn for readers. Does anyone know of a web site that posts the actual cash value of the Powerball jackpot after the draw? There are a number of Powerball state web sites that posts the annuity value after the draw but I cannot find a site that posts the cash value. Thanks.

Good luck,
Jake

United States
Member #2338
September 17, 2003
2063 Posts
Offline
 Posted: November 7, 2005, 2:52 pm - IP Logged

The cash value for the recent large Powerball was 164.4 million. It didn't go up from sales thanks to the idiotic cap.

While it would be nice if MM could finally break 300 the odds of it breaking 400 must be lower than that unless sales take off. At this point I can't see it lasting more than 3 rollovers.

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 7, 2005, 3:08 pm - IP Logged

Yes, the record jump is as you stated Prob988 - \$133 million in May 2000.

Unfortunately we cannot know for certain if that jump was entirely based on sales or was it partially based on the poor estimate of the \$230 million jackpot. I have noticed that Mega Millions (and Big Game) never adjusted their estimates after the draw to the actual value based upon sales. Therefore the \$230 million jackpot may have grown to say \$263 million by the time of the draw. Under this case, the jump was only \$100 million.

Powerball on the other hand always states the actual jackpot value after the draw even if the jackpot was not won. A questinn for readers. Does anyone know of a web site that posts the actual cash value of the Powerball jackpot after the draw? There are a number of Powerball state web sites that posts the annuity value after the draw but I cannot find a site that posts the cash value. Thanks.

This site gives the sales figures for those jackpots, Jake:  http://www.lottoreport.com/mmsales1.htm

They were as follows:

\$195,576,985
\$122,470,608
\$67,249,897
\$49,051,930
\$23,336,032
\$16,770,708
\$13,388,353
\$11,524,249
\$9,507,698
\$8,569,000
\$7,291,990
\$6,932,334
\$5,709,195
\$5,642,696
\$4,660,016
\$5,003,855
\$4,081,343
\$4,418,577
\$3,829,707

The total was \$565,015,173, with 34.61% represented by the last drawing.  36.6% of the final total jackpot (and I have used the announced jackpot of \$363M, not the figure given in lottoreport - lottoreport only reports the "advertised" jackpots) came from the last jump.  Therefore you are probably correct, the real jackpot jump was somewhat lower.

You can calculate cash value from the figures the lottoreport site, using another page on the lottoreport site:  http://www.lottoreport.com/salescomparison.htm#pb2005.  If you sum the sales for a drawing, and multiply by 0.318, you will get the cash value.  The spreadsheet where I keep my records for megamillions automatically does it.  For your reference, here are the cash values of the recent run of rollovers, as so determined from my spreadsheet.

 \$113,089,857.45 \$97,939,964.29 \$85,939,473.40 \$74,631,212.53 \$64,908,822.25 \$54,573,259.95 \$46,215,124.18 \$38,822,651.15 \$32,482,274.08 \$25,807,824.52 \$20,060,962.18 \$14,420,071.19 \$9,612,621.82 \$4,664,967.33

The annuity values are determined by prevailing market conditions on the purchase of annuities.    Since these can vary, the annuity jackpots are somewhat less indicative of the real state of affairs in terms of sales and the cash put into the prize by the lottery to make these purchases.

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 9, 2005, 10:00 am - IP Logged

The Megamillions has rolled over again, and the cash value has risen.  Based on sales, the cash value reached  \$131.5 and is expected to rise to \$153.4, the annuity value from \$165M to \$192. This means that the lottery expects to sell about \$70.2M tickets.  From these values I estimate the probability of various numbers of winners as follows:

 0 67.06% 1 26.79% 2 5.35% 3 0.71% 4 0.07% 5 0.01%

My long term prediction developed above was off by about 5%, which I don't consider all that bad, given that it is only a model.  That model was developed using data from the last run.  If I substitute the data from the last three draws on this run to generate the exponential function parameters I get the following equation:  S = \$2,416,000*exp(0.212*N) where N is the drawing number, including the starting draw, and S is the predicted sales for the draw.  This function gives a long range probability outlook that runs like this:

Annuity                            Cash Value              Prob of Rollover        Prob of Rollover

 \$654,794,076.40 \$381,235,662 30.50% 1.14% \$540,828,466.19 \$314,882,351 38.28% 3.75% \$448,663,818.45 \$261,222,045 46.00% 9.79% \$374,129,735.07 \$217,826,646 53.36% 21.29% \$313,853,591.00 \$182,732,535 60.17% 39.90% \$265,107,924.87 \$154,351,725 66.31% 66.31%

Most likely the actual situation is somewhat between these values and the values from the equation derived in my previous post with different parameters for the exponential equation.  The former equation, updated a little to reflect the data from this run with respect to existing sales, gives long range probabilities looking like this:

 \$729,669,651.32 \$424,829,886 16.64% 0.52% \$557,571,030.73 \$324,630,245 27.07% 3.15% \$432,167,639.89 \$251,617,604 38.59% 11.63% \$340,789,740.51 \$198,415,360 49.97% 30.14% \$274,205,253.13 \$159,648,392 60.32% 60.32%

I suspect that the lower model, while somewhat further off the mark for the current drawing, may be better in the long run when, if there are many rollovers, a stronger case of lotto fever takes hold.  Specifically I believe that there is a 30-40% of two more rollovers, which is quite signifcant under the circumstances, and that if this happens the annuity jackpot will be in the neighborhood of \$400M annuity.  I believe the next jackpot, if it occurs, will have a cash value greater than Jack Whitaker's prize.    I also believe that any jackpot that is won, will still most probably have just one winning ticket.

Both equations give roughly a 10% chance of breaking \$500M annuity, though the first equation takes longer to do it.

BILOXI, MISSISSIPPI
United States
Member #19651
August 3, 2005
621 Posts
Offline
 Posted: November 9, 2005, 2:22 pm - IP Logged

you the man             i always love your work    now could you -give odds on next state based on sales  nj ny ca il ga    or just call it    alt ga

BILOXI, MISSISSIPPI
United States
Member #19651
August 3, 2005
621 Posts
Offline
 Posted: November 12, 2005, 12:30 am - IP Logged

will it be 55%sat with roll over and est for roll over jack pot for fri thanksss

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 12, 2005, 5:53 am - IP Logged

The Megamillions has rolled over again, and the cash value has risen.  Based on sales, the cash value reached  \$131.5 and is expected to rise to \$153.9, the annuity value from \$262 to \$310. This means that the lottery expects to sell about \$90M tickets.  From these values I estimate the probability of various numbers of winners as follows:

 0 59.92% 1 30.69% 2 7.86% 3 1.34% 4 0.17% 5 0.02%

We now see that the most probable outcome, based on this data, is yet another rollover.  The second most probable outcome, by far, is a single winner.  The new jackpot will exceed Jack Whitaker's pot, and thus if a single winner is obtained, it will be a true record single winner jackpot on a cash basis, unlike the recent \$340 annuity jackpot for powerball which was inflated via an alternate type of annuity.

The first long term probability equation given in my last post of this type, predicted within 1% the size of the current jackpot.  This was pretty good, I think.  The long term probabilities, given using the exponential model generated from the last three data points are as follows:

Annuity Value                  Cash Value                    Single draw              All draw

Rollover Prop            Rollover

 \$617,878,872.37 \$363,751,272 34.32% 2.34% \$516,362,430.32 \$303,987,560 41.50% 6.81% \$432,883,802.91 \$254,842,884 48.52% 16.41% \$364,237,964.90 \$214,430,415 55.17% 33.83% \$307,789,371.76 \$181,198,582 61.32% 61.32%

Since this run and the previous run both now have run the same number of draws, 16, it is now possible to generate an exponential equation based on the average sales for both runs.  Using 4 such data points, that model gives the following figures.

Annuity Value                  Cash Value                    Single draw              All draw

Rollover Prop            Rollover

 \$770,379,111.36 \$448,531,838 19.06% 0.51% \$611,309,616.55 \$355,918,043 27.83% 2.69% \$488,559,145.81 \$284,449,992 37.27% 9.67% \$393,835,276.31 \$229,299,650 46.69% 25.94% \$320,738,925.65 \$186,741,330 55.56% 55.56%

Comparing the two model equations we see that, according to them, there is 16-26% chance of reaching an annuity jackpot of around \$500M.  There is a 7-10% chance of reaching \$600M.  The first model accomplishes these things in one more drawing.

BILOXI, MISSISSIPPI
United States
Member #19651
August 3, 2005
621 Posts
Offline
 Posted: November 12, 2005, 3:46 pm - IP Logged

hay could we take the cash    put it in  past \$ VALUE  and get an idea  what jacks jack-pot was this one would be 375 ? or much less  thanks {MY  WIFE THINKS YOU DO REPORTS FOR WALLSTREET }

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 17, 2005, 9:16 am - IP Logged

The MegaMillions jackpot has been won.  We can expect that the reset jackpot will sell around 14.5 million tickets.  The probability of various numbers of winners is as is given in the following table:

 0 92.34% 1 7.36% 2 0.29% 3 0.01%

Let's do a post mortem on the last \$315M jackpot.    The actual sales for this drawing was \$104,813,365, giving a cash jackpot of \$187.1M, an all time record.  The final probabilities for numbers of winners is given in the following table.

 0 55.07% 1 32.85% 2 9.80% 3 1.95% 4 0.29% 5 0.03%

Thus a rollover was more probable than a single winner, so having a single winner beat the odds, but not by all that much.

I have worked on a long term model for the growth of lottery sales using the average sales at each drawing point of the last two sets of rolls, that which went to \$250M and that which went to \$315M.  The synthetic equation I generated is as follows:

S = \$9,748,346.62 * exp(0.118N)

Here S is the sales for a particular drawing number, and N is the number of the drawing, including the first reset drawing which is 1.  The results of this equation to not exactly match the behavoir of the lottery for every drawing, but I think it's not a bad approximation.

Here is a table comparing historical results and what this equation predicts:

 \$526,047,630.15 \$306,276,620 51.28% 0.42% \$461,945,474.94 \$268,954,921 55.25% 0.81% \$405,005,754.56 \$235,803,350 59.04% 1.47% \$354,428,176.51 \$206,355,961 62.62% 2.49% \$315 \$309,501,868.77 \$180,198,866 65.98% 3.98% \$250 \$262 \$269,595,388.37 \$156,964,426 69.12% 6.03% \$200 \$225 \$234,147,846.42 \$136,326,079 72.03% 8.72% \$172 \$192 \$202,661,024.79 \$117,993,752 74.72% 12.10% \$150 \$165 \$174,692,373.60 \$101,709,782 77.19% 16.20% \$131 \$147 \$149,848,791.14 \$87,245,296 79.46% 20.98% \$111 \$128 \$127,781,098.82 \$74,396,995 81.53% 26.41% \$97 \$108 \$108,179,133.39 \$62,984,295 83.41% 32.39% \$84 \$90 \$90,767,387.60 \$52,846,790 85.12% 38.84% \$72 \$77 \$75,301,137.94 \$43,841,996 86.66% 45.63% \$61 \$65 \$61,563,004.97 \$35,843,350 88.06% 52.65% \$51 \$53 \$49,359,898.12 \$28,738,430 89.32% 59.79% \$41 \$42 \$38,520,301.70 \$22,427,376 90.45% 66.94% \$32 \$32 \$28,891,864.29 \$16,821,485 91.47% 74.00% \$23 \$23 \$20,339,257.41 \$11,841,968 92.39% 80.90% \$15 \$15 \$12,742,273.46 \$7,418,835 93.21% 87.57% \$12 \$12 \$5,994,136.20 \$3,489,919 93.95% 93.95%

The first column shows the historical values for the jackpots for the \$250M run, the second, the historical values for the jackpots for the \$315M run.  The third column shows what the equation predicts as an annuity value from the calculated cash value.  The fourth column is the cash value calculated from the sales predicted by the equation given above.  The fifth column gives the probability of a rollover for a jackpot having sales predicted by the equation.  The fifth column contains the probability that that jackpot and all previous jackpots in the series will rollover.    (The calculated sales figures are not shown themselves.)

One can see that for several regions - not including the earliest drawings where lottery policies inflate the prize - the equation represents a decent match with what actually occurs.  This gives one a rough idea of how likely various jackpot sizes are at the time of a reset.  It is just an approximation, of course, but it is a reasonable one I think.  This fit is very good in the middle drawings, and not all that bad in the higher drawings.

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 20, 2005, 11:39 pm - IP Logged

The Megamillions has rolled over again, and the cash value has risen.  Based on sales, the cash value reached 5.4 million and is expected to rise to \$9.4, the annuity value from \$12 to \$16M. This means that the lottery expects to sell about \$12.6M tickets.  From these values I estimate the probability of various numbers of winners as follows:

The probability of exceeding the large last prize, 315 M, is about 3.7%, of matching it, around 6.7%

New Jersey
United States
Member #21206
September 4, 2005
949 Posts
Offline
 Posted: November 20, 2005, 11:41 pm - IP Logged

The Megamillions has rolled over again, and the cash value has risen.  Based on sales, the cash value reached 5.4 million and is expected to rise to \$9.4, the annuity value from \$12 to \$16M. This means that the lottery expects to sell about \$12.6M tickets.  From these values I estimate the probability of various numbers of winners as follows:

 0 93.08% 1 6.67% 2 0.24% 3 0.01%

The probability of exceeding the large last prize, 315 M, is about 3.7%, of matching it, around 6.7%

 Page 3 of 24