Subscribe to DSC Newsletter

Invented by a data scientist: the first anti-scam

By anti-scam, I mean a scam where the victim makes more money by participating in the scam.

This scam works as follows:

  • You send money to the data scientist - AKA the scammer - and he purchases lottery tickets with your money, 
  • He claims to be using a secret data science algorithm to pick up lottery numbers to increase odds of winning (he won't tell you the numbers he chose, for fear that his secret algorithm might be reverse-engineered) 
  • The scheme truly provides returns / odds of winning far above what lottery typically offers

The scam is offered by a very respected data scientist (you need to create trust), and indeed, despite being a scam, it works as advertised - even better - for the victim. In short, the victim is happy.

So where's the trick? What's the secret algorithm? It's actually very simple:

  • The data scientist keeps 10% of the money received to pay himself a salary or commission.
  • He uses 30% of the money to play in selected lotteries (with no secret algorithm - that's the scam part)
  • He redistributes 60% of the money, semi-randomly, straight to the participants (some receive less than they paid, some receive more)

A typical lottery (see picture below) returns 50% of the money to the player, on average. So the data scientist should expect to recover 50% of the 30% he invested in the lottery - probably more since he's doing some kind of lottery arbitrage to boost the return for the participants, by only playing the most profitable (least loss) lotteries.

In short, the data scientist (the "scammer") returns (50% * 30% + 60%) = 75% of the money to the player. This is much better than lottery (50%). In addition, the data scientist makes a lot of moderately small winners rather than very few spectacular winners. This creates more happy victims and lower the risk of complaints.

Question: Do you know other scams where the victim truly gets a benefit? Is this really illegal or not? My wife told me it is a scam, but I'm not convinced it's one. Also, you could replace "lottery" by "Wall Street" and create a new anti-scam. Or, instead of playing 30%, he could not play at all, keeping all the money as cash, and redistribute 90% to the participant, rather than an expected 75%.

Related articles:

Views: 2988


You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

Comment by Vincent Granville on April 30, 2013 at 9:48am

Another anti-scam idea is about healthcare. For those hit by the Obamacare health penalty (e.g. small business owners) and who don't want health insurance because they use a much more efficient system to take care of their health, you offer fake health insurance for $20/month. All participants know full well the insurance is fake (and for instance, if you need treatment, maybe the only available doctors are in North Korea). It is offered to selected people, not to the public at large. It is set up abroad by a foreign company, and participants can claim on their tax return that they have health insurance from a foreign company. Better, this insurance company is set up by people who will never use it, and complies with all laws to qualify as an official health insurance company. By participating, you can save $10,000 per year in ridiculous fees.

Comment by Vincent Granville on April 26, 2013 at 3:45am

The 50% is correct because it represents many small losses with very few (<1.5%) very large gains. It does not represents a chance of winning, but instead the average loss computed on many purchases. 

Comment by Eduan Nel on April 26, 2013 at 3:24am

Hi Vincent, yes thats the point, the real odd's of winning playing the lottery is much less than 50%. Let say its 1.5% for example, then the formula should have been 1.5%*30% + 60% = 60.45% . So the scam will still work because of the very low odds of winning on the lottery.(1.5%) Reg, Eduan

Comment by Vincent Granville on April 25, 2013 at 1:01pm

Hi Eduan - the 50% is the number advertised by the lottery itself. While much less than 50% of people get any money back, a handful get tons of money. That's how the 50% are re-distributed: 40% goes to 0.001% and the remaining 10% goes to maybe 1.5% of all players. In short, 98.5% of players get nothing.   

Comment by Eduan Nel on April 25, 2013 at 7:17am

Question, how do you get to the 50% of money paid to the player ? Think about it, it can never be - the real odds for you as a player is much less than 50%. I't will be close to 1 in 55. The 50% is allocated to prizes money (not players)  and not equally distributed.

Comment by Mirko Krivanek on March 30, 2013 at 9:32pm

Here's another, weird anti-scam: a web site that sells fake Twitter followers ($10 for 1,000 followers) to make you appear more popular than you really are, supposedly to boost your sales. They tell you upfront that these followers are fake. They also sell Facebook "likes", though they claim that they are from "real" Facebook profiles.

You need to be a data scientist to sell fake "likes" and "followers" on a massive scale, because you need to be smarter than Facebook and Twitter spam detection algorithms (designed by data scientists paid $133k yer year). However these spam detection algorithms are currently quite weak and easy to crack: maybe on purpose - after all shareholders don't care if the users are real or fake as long as there is growth and everybody believes in the growth story, that's how bubbles work. So maybe a kid in Russia can beat these algorithms without being bothered, and make money by selling this fake traffic.

The easiest way to detect fake "likes" is by looking at the number of relevant comments: if a Facebook (or any) post has 5,000 "likes" and either 0 comment or 20 comments that are just variations of "this is a great post", then you bet the "likes" and "comments" are mass-produced by a robot or an army of zombies - they are fake.

Comment by Sam Kaplan on March 29, 2013 at 5:07am

First, the premise is wrong.  The "victim" would do better putting the money in a 1% interest CD, as on average he/she would receive 1.01 times the initial amount rather than .75 on average.  So, the victim is better off not participating.

Second, it is fraudulent in that the "anti-scammer" is not doing what is claimed to be doing.  By contrast the lotteries that Vincent describes are quite open about how much is distributed as often the proceeds are supposed to go for education or some other public purpose when they are government-run.

Third, it is a lottery, albeit a secondary lottery (using someone else's).  

Fourth, it is not the percent of money distributed but how the distribution is accomplished and what the distribution is.  Consider a chain letter where the first person on the list is not the anti-scammer, who claims zero dollars, but some other person or institution--it would still be illegal in the US.  

Fifth, if you did it with Wall Street, you would go to jail as did these folks:

Comment by Vincent Granville on March 29, 2013 at 12:12am

If produced on a large scale, the big losers would be lottery operators (States?) since much less money would actually be "played". Everybody else would indeed win. Not sure how lottery operators could successfully sue. Part of the issue is that the lottery redistributes very little money (50%), this could not happen if they redistributed 90%, like the casinos. The only way to prevent arbitragers from taking advantage of the opportunity is very heavy regulation, as in healthcare: you don't have middle men selling in US for $2 a drug bought in India for $1, and sold in US by big pharma for $4, because of heavy regulations, enforcement and penalties. Arbitragers in these schemes get heavier prison sentences than murderers, despite the fact that they committed a victim-less crime.

While this seems unfair, at least lottery can be seen as the tax on the innumerate, thus easily avoided by analytic people. If these lottery arbitragers were to be successful, we would all end up paying more taxes. 

Follow Us

On Data Science Central

On DataViz

On Hadoop

© 2017 is a subsidiary and dedicated channel of Data Science Central LLC   Powered by

Badges  |  Report an Issue  |  Terms of Service