# Flipping Pennies Odds

I flipped a penny 50 times to see what happened.

My set of results were:

HHTTHTHTTT
HTTTTTHHTT
TTHTTTHHTH
HTHTTTTHHH
THHHHHTHHH

So I had 26/50 tails results, almost exactly half. I would expect this to follow through if I flipped the coin 100 or 1,000 times.

However you need to let the coin land and not spin on the table. Pennies really do have different weights on each side because of the designs on them!

http://www.sciencenewsforkids.org/pages/puzzlezone/muse/muse0403.asp

So if you're going for true randomness pick cards out of a deck and don't flip coins that might roll :)

If you want to see an awesome movie that talks about coin flipping and probability, watch "Rosencrantz and Guildenstern are dead". It begins with them walking along flipping coins and talking about the probabilities. They hit 92 heads in a row.

Probabilities are all around us. Certainly anyone who gambles needs to understand probabilities. When my boyfriend used to shoot darts in keno bars, I would sit and watch the keno numbers for about an hour and keep track of any which hadn't come up at all. Then I would start progressively betting on them (i.e. \$1 the first time, \$2 the second time, \$4 the third time) that way whenever I won I would make up the money I'd put into it already and get some money out. In theory this would work well - but in practice the Keno system has computer programs watching people's bids and adjusting the results of future games in order to balance the overall payout of the system. I.e. it wasn't random at all, and bartenders told me they'd seen people doing this very thing and suddenly the Keno system would go 50 times in a row without ever showing a given number. Then the person would stop their betting and poof the number would go back to its normal ratio. So I gave up that project :)

We all watch the weather and determine the probability of snow or rain to make our plans for the day. We take into account the probability of a fire in the placement of our fire extinguishers. We account for the probability of traffic issues in building 'buffer time' into our departure time when we're trying to go somewhere.

In my website business I have to figure out the probability that I'll have a site failure to figure out how frequently I want to do backups. I don't want to back up every five minutes! But I also don't want to back up every 5 years either. If my site was failing every day I might back up daily though. (And I'd get a new hosting facility). If there's a high probability that a given lesson in our plagiarism training course is going to confuse our trainees, I'll add on a quiz to help them further understand and remember that lesson. If there's a low probability of confusion, I'll skip the quiz and let them move on to the next lesson.

(someone asked me about the Keno situation and if that was legal for them to do)

To be fair, a reason that gambling organizations run the way they do is that they are aiming for a certain payout level. Many countries regulate the payout level of casinos to make sure they are fair. So let's say just as a wild example that a given casino is regulated and is required to have a payout level of 99%. That is, for every \$100 they bring in, they are required to give \$99 of it back to its guests. They keep \$1 out of every \$100 gambled.

Now say this casino just let wild chance run rampant in their casino, and in a given month there was a streak of situations which meant they actually had a payout level of 95%. They ended up keeping \$5 out of every \$100 gambled. It can happen, of course, depending on what people choose to bet on! Now the casino would be in huge trouble for keeping too much money. It wasn't their "fault". It was just the odds. But they are required - to be a fair casino - to pay out more than that.

So what they do is balance all the games and systems so the average payout level is properly met.

In terms of slot machines, from wikipedia:

"Slot machines are typically programmed to pay out as winnings 82% to 98% of the money that is wagered by players. This is known as the "theoretical payout percentage" or RTP, "return to player". The minimum theoretical payout percentage varies among jurisdictions and is typically established by law or regulation. For example, the minimum payout in Nevada is 75%, and in New Jersey, 78%."

Every slot machine I've seen states right on it what its required payout level is. So it HAS to give out that much. How could it do that if it was left to random chance? There would always be a chance - a remote one maybe but a chance - that it could fail to meet that level. It absolutely has to have some artificial intelligence watching it and making sure it meets its required levels.

Statistics Basics