That's... Really, seriously, that's not how it works. Going to the binary example, I will do my best to break it down. You have a .5 probability of heads *or* tails on a coin toss, no matter how many times you have flipped it, no matter what comes up prior to the flip. So, we want to calculate the odds of flipping heads three times in a row. As such, we multiply .5 by .5 by .5. This results in a .125 probability, or a 12.5% chance. With me so far? Okay, now let's review the possible combinations of those three coin flips: Heads, heads, heads Heads, tails, tails Heads, heads, tails Heads, tails, heads Tails, heads, heads Tails, tails, heads Tails, heads, tails Tails, tails, tails That's 8 potential outcomes. Each of these outcomes is equally possible. If you can't see that, this next step should make it clearer. .125 multiplied by 8 equals... 1. As such, we have clarified each potential outcome and correctly identified the probability for each possible combination. They are each equally probable. As such, each time we flip the coin, regardless of prior results or the number of times we have flipped the coin, there is a 50% chance of heads and a 50% chance of tails [b]because each outcome is equally possible[/b]. It's math. It works out.

But what everyone else is saying is that there is the same chance to get 4 other items in the 4 weeks. For example 2 heads and 1 tail. They just look at it like .5*.5*.5 But what can really occur is HHT HTH THH which is 3/8 (which is more than 1/8) Therefore getting other combinations of the items is more likely than 4 in a row

That's not what happens. I fully understand what is being said but that's like saying combinations that don't produce a specific pattern have a zero percent chance of happening because they didn't happen. It's manipulating results to support a hypothesis. Each combination has an equal chance of occurring. Whether that result presents a specific combination or a combination of all items is not important. A series of 10 heads in a row in a sampling of 100 coin tosses means nothing. The distribution of heads to tails is going to approach 50% the more attempts that are made. The same can happen here after 100 weeks. The probability of any one item being on the vendor on any given week is 16.67%. You can't point at a six item list after 12 or so picks and say that four results in a row are statistically significant in any way, shape, or form.

That's not what I'm saying, I'm just saying that this guys calculations and assumption are correct. Not once have I referred to weeks previous the last 4

All I'm trying to say is that getting 4 results that are the same in a batch of results is less likely than other combinations Eg. 3 same 1 different Or 4 different