Estimating how many lottery tickets were sold

About scratch lottery

These numbers are available:
- How many tickets were printed
- how many winning tickets were printed for every prize levels
- how many winning tickets were claimed for each prize level

Assuming that the people who scratched winning tickets have claimed the prizes, is there a way to estimate how many tickets were sold within a certain interval of certainty by looking at the number of claimed prizes?

I am not asking for someone to calculate this, I just want to know what I should search for. I'm pretty good at math but haven't done prob and stats for a long time and I don't remember the name of the techniques I should be reading up on.

Thx.

What are you going to do?
Drive the entire state buying up all remaining tickets when you identify a positive EV situation?

Quote: Wonko33

Assuming that the people who scratched winning tickets have claimed the prizes, is there a way to estimate how many tickets were sold within a certain interval of certainty by looking at the number of claimed prizes?

I am not asking for someone to calculate this, I just want to know what I should search for. I'm pretty good at math but haven't done prob and stats for a long time and I don't remember the name of the techniques I should be reading up on.

Thx.

I would say so, but I'd suggest you'd only really need to look at the lowest (most common) prize level, unless you happen to know the overall probability of winning, then look at the total number of winning tickets.

From there, a really simple way would be to goof off with the Vassar Stats binomial Distribution calculator. What you would do is put the overall probability of winning, say it's 0.25, then you put in the number of winning tickets claimed, finally, your Variable is how many tickets you want to be sold.

http://vassarstats.net/binomialX.html

For example, I decided the overall probability is 0.25, and now I want to see how likely it is that I have 10000 (or more) winners in 39000 tickets sold. It tells me the probability of that is 0.175%, so there is a 99.825% chance it'd have taken more tickets sold.

If I use 40500 sold, then I get a 92.5275% probability of there being 10000 or more winning tickets. So, there is only a 7.4725% chance I sold less than that, based on the number of winners. Anyway, you could go trial and error with that and arrive at any percentage of confidence you like. It works best for large sample sizes.

Thx Mission, I'll go check on Vassar

Quote: MaxPen

What are you going to do?
Drive the entire state buying up all remaining tickets when you identify a positive EV situation?

Yes, I'll find a small state, I'm thinking Rhode Island, and I'll do that. Thank you for taking the time to answer, even if you had nothing to say.

Quote: MaxPen

What are you going to do?
Drive the entire state buying up all remaining tickets when you identify a positive EV situation?

This is exactly how you AP pull-tabs or jar tickets. And it's even easier because there's no driving involved. Just find a ticket set with more prizes left than the cost of the remaining tickets. If you can properly eyeball the number of tickets left, you can get the edge. It's called a buyout.

Ask her..............

http://www.philly.com/philly/news/nation_world/Lotterys_luckiest_woman_Joan_Ginther_bet_flabbergasting_sums_on_scratch-offs.html

Quote: wellwellwell

Ask her..............

http://www.philly.com/philly/news/nation_world/Lotterys_luckiest_woman_Joan_Ginther_bet_flabbergasting_sums_on_scratch-offs.html

I know, she's why I'm looking into it. I'm trying to see how improbable it is that she did it that way . I was looking into fun exercises to get back into stats and that is one I came upon. I don't think that's all she did it, probably part of it though. I read that retailers use to complain about getting "stale" bundles with very few winners, so the printing company doesn't really randomize the winners anymore, they have a way of spreading out the winners. Rumor is she had inside info about the distribution centers too.

Probably a waste of time, but getting the rust off the old stat skills is better than watching reality TV ;)

I would urge you to keep in mind that thousands and thousands of winning tickets go unclaimed every year in virtually every State, so even with the information...I just don't think it's reliable enough.

Further, I don't know if you're just looking for overall +ER, or +ER to go for a long-shot, but even if a long shot win is your goal, there are better ways to do it with more reliable (read: exact) numbers.

Quote: MathExtremist

This is exactly how you AP pull-tabs or jar tickets. And it's even easier because there's no driving involved. Just find a ticket set with more prizes left than the cost of the remaining tickets. If you can properly eyeball the number of tickets left, you can get the edge. It's called a buyout.

The pull-tabs I have seen , the containers are hidden so you can't tell how many were sold unless you are friendly with the staff

Quote: Mission146

I would urge you to keep in mind that thousands and thousands of winning tickets go unclaimed every year in virtually every State, so even with the information...I just don't think it's reliable enough.

Further, I don't know if you're just looking for overall +ER, or +ER to go for a long-shot, but even if a long shot win is your goal, there are better ways to do it with more reliable (read: exact) numbers.

it is more of a thought exercise, and yes I know many tickets go unclaimed, especially the low denomination wins. Since those are the ones in greater quantity I'm guessing it can really mess up the estimate I think. Also I thought it was weird that this data is posted, so I want to see how irrelevant having this info is, which is probably why they do post it.

Quote: MathExtremist

This is exactly how you AP pull-tabs or jar tickets. And it's even easier because there's no driving involved. Just find a ticket set with more prizes left than the cost of the remaining tickets. If you can properly eyeball the number of tickets left, you can get the edge. It's called a buyout.

I think you are getting into a paradox.

The OP try to use the number of prize claimed and the probability to work out the ticket sold, then you try to use that figure (ticket sold) to find "a ticket set with more prizes left than the cost of the remaining tickets."

By definition I think the prizes will be exactly at par with the cost.

Quote: andysif

By definition I think the prizes will be exactly at par with the cost.

I disagree, but with that said, I also don't think you're going to find much of an advantage, either. If you do find an advantage, it would involve hitting some sort of long-shot, so your, "Drop," between good hits is going to be nasty.

The first thing to remember is that these scratch-offs routinely pay as low as 50% ER, so that's objectively terrible. However, much like something like Powerball but not quite to that extent, a good portion of this return is going to be concentrated in the higher paying results. The result is that you're looking for a disproportionately low amount of the larger returns claimed, and that's where your ER is going to show some improvement. Will it be enough to overcome the House Edge? I imagine that it will sometimes, rarely...and you'd also need to be pretty conservative to account for unclaimed tickets, as well.

Are you going to be the one that hits it? Well, no, probably not, but it's still no different than playing some sort of linked Slot/VP Progressive at +ER, or something along those lines. You might not be the one that hits either of those things, either. I guess if you only play with an advantage you should win in the long run, but Hell, you could have a situation where two of the top prizes have been sold, but are simply unclaimed as of yet, and POOF, +ER gone.

I think that many of the scratch off winners that have happened over the long run (and there haven't been many) involved something like a GWAE interview I heard once in which it was determined that there was some kind of pattern because the tickets were made in an inferior way, or you could tell if it would be a winner without scratching, something along those lines.

Quote: Wonko33

it is more of a thought exercise, and yes I know many tickets go unclaimed, especially the low denomination wins. Since those are the ones in greater quantity I'm guessing it can really mess up the estimate I think. Also I thought it was weird that this data is posted, so I want to see how irrelevant having this info is, which is probably why they do post it.

It could mess up the estimate, and the more ER comes from those lower-denomination wins, the more that could skew the estimate. I know some Religious scratch-off enthusiasts (read: idiots) that will keep all of the $2 or less wins until the end of every month and then cash them all at once. By, 'Some,' I guess I mean, 'One,' but I imagine there are others who do something along those lines.

I'll tell you what I said to him, "You've been a good friend to me, but if I'm ever behind you in the gas station when I'm running right on time to get to work and you are in front of me having the clerk scan fifty tickets, I'll not speak to you from that day forward." Fortunately, they have the thing where you can just scan it now and get credits, and maybe it even dispenses cash. I don't play the scratch-offs to know, I just know that it definitely gives credits for more tickets.

Mission I'd go as far to say most of these would
Be a gift at 50er it's more like 33.33 and even 25. From what I know
In pa on the 30 dollar tickets odds stated on the back are 1/2.93

For return or for the ticket to have some kind of return? What scratch-offs I've seen have almost always (if not always) stated the overall odds of winning, so maybe that is what you were seeing. I didn't mean 50% of tickets have some kind of win, I meant the overall % return of the game.

According to state mandate the minimum the instant lottery in pa must return is set at 30 percent however is 2011 a temp injunction was granted to lower that to 27 percent - it has since be raised back to 30 percent

Man, that's ugly!

That being said- lottery does lost overall returns on instants at 55 percent roughly for last year. I'm sure the higher tickets account for that and the lower rixkets are the ones that get hurt the most

I personally don't really like scratch off tickets too much as they can be huge money suckers. I have bought 12 consecutive tickets of the same name and gotten only a quarter of my money back. For example, I bought scratch off tickets 20 21 22 23 24 25 26 27 28 29 30 31 32 on a $1 scratch off ticket called Quick Bucks, and only numbers 24, 30, and 25 were winners, and each of them was only worth $1.

Quote: Dean

I personally don't really like scratch off tickets too much as they can be huge money suckers. I have bought 12 consecutive tickets of the same name and gotten only a quarter of my money back. For example, I bought scratch off tickets 20 21 22 23 24 25 26 27 28 29 30 31 32 on a $1 scratch off ticket called Quick Bucks, and only numbers 24, 30, and 25 were winners, and each of them was only worth $1.

What's the problem? You ran pretty good!

Ontario sets their odds between 50% and 70% depending on the value of the ticket. There is virtually no way to get an edge on these. They also publish the number of unclaimed prizes down to the $100 level which allows you to estimate how many tickets have sold by taking the total number of tickets x the number of prizes claimed x the odds of winning that prize.

The way I look at it the top prizes only compose a few percent (perhaps at most 10%) of the total prize pool. Even if none of these top prizes are taken, you would need to get into a situation where the number of unclaimed winners increases the edge to over 100%.

This is relatively calculable.

There are factors to overcome:

(1) Unclaimed winners - tickets are sold until the prizes are claimed, not sold. That allows the lottery to not reduce that number of unclaimed prizes until the prize has been claimed and it inflates the tickets available.

(2) Variance. You would need to have a huge bankroll to pull a winner and even with this there are many times of failure.

Thx for all the responses,

All that you guys are saying is just reinforcing my hunch that that Texas woman could not have hit those jackpots just by calculating when it was the right time to buy, she must have had more info about the distribution of winners or something.

As for the reply concerning a pattern to find winners, the person who publicized it is MOHAN SRIVASTAVA, a statistician. I've heard interviews with him, he's pretty interesting fellow, his technique for spotting winners on certain types of scratch lotto is very simple.
Problem is you have to be able to look at the tickets for a few minutes, so you need a friendly vendor, own a 7-11 or work the night shift at one :)- also now many places have the vending machine sale points where you can't see the ticket

Quote: Wonko33

The pull-tabs I have seen , the containers are hidden so you can't tell how many were sold unless you are friendly with the staff

In Washington State?
From WAC 230-14-055

Quote:

(b) Sell all pull-tabs from a dispenser we approved or a clear container. Pull-tabs sold from a container must be visible to players so players are able to estimate the number of chances remaining in the series;

If you want formatted data on scratch tickets from WA, go here and just change the game id .

Quote: Wonko33

Yes, I'll find a small state, I'm thinking Rhode Island, and I'll do that. Thank you for taking the time to answer, even if you had nothing to say.

Whatever, the question was asked in return because the original question seemed a little wonky to me.