i totally agree with what she said, so i'm going to be recording to see if there is a group of 5 or 6 reds or blues somewhere.

i will post pics for this in this thread.

Yes, and show 350, conor's game



There is a group of reds in the west wing! Thanh is right so far. Let's see if theres anything on sunday!

But how does a group of red or blue boxes help you know which boxes to pick?

You might well get a pattern that emerges by the important part of the game, allowing you to possibly hit blues with more consistency. It doesn't actually change the odds, but it does make sense that you get patterns.

This is going to be darned interesting to keep track of though.

Yep, there was 5 reds in a row as well!?

its amazing how this has worked over 60 times

60 times!?

Shame no-one used it when it worked then.

worked??? its still working

I wonder if it'll work on Sunday, will you be writing it down Aaron?

Indeed.

The principle here is just the same as something like Mally's Law, in so much as it looks at what events have already happened, compares them with the 'average' and (wrongly) concludes that you can make an accurate prediction based on that.

Compare and contrast these flawed arguments...

1. The quarter-million hasn't made it to the table for 100 games. On average it'll be there once every 22 games. Therefore it's very likely that it'll be there soon.

2. There has been a cluster of 3 reds so far on the West Wing. On average, you'd expect a cluster of 4 or 5 of the same colour somewhere. Therefore it's very likely that the next box along also contains a red amount.

Both of those arguments are fallacious. Boxes do not have any awareness of which colours are in boxes near them, just like they have no awareness of what has happened in prior games.

These clusters may be interesting to keep track of, but if anybody thinks they can be used to give the player a mathematical advantage, then they are, to be frank, quite mistaken. (As Thanh found out to her cost this week...)

could this "system" work in the order of the numbers. for example could boxes 6-10 each have the same colour

thanks for the idea tr4962! i wil simulate that tomorrow aswell!

That would give exactly the same result. Namely, that it is very likely that there'll be a cluster of the same colour somewhere, but that it'd be impossible to accurately predict where.

actually, i did record the numbers, so i will type them in and tell you whether there is something about the numbers

I think this is interesting to look back on and see if there are clusters, but you can't really use it in your game as a system, because it won't work every time.

I'll be interested to see how often it does work, though!

nope, no pattern

I've worked out the following:

*deleted*
*did it wrong, might try and redo for a 3rd time to get correct*

i dont think you get it. They have to be next to each other, like the grop of 5 reds

oh right, so straddling the bit between the wings doesn't count as next to each other?

I will try and adjust the odds.

Not that I wish to be the bearer of bad tidings, but you have messed up somewhere. The colours in the boxes are not independent of each other. [Just like the probability of an all-blue first round is not (1/2)^5. It is (11/22)x(10/21)x(9/20)x(8/19)x(7/18).]

The actual probability of getting a cluster of 11 (allowing wing-straddling) is a mere 0.00312%, roughly. I doubt it's happened yet, and I'd be very surprised if we ever do get to see it.