..but we don't have the topic 'whiney loser who doesn't know his probabilities'..yet...

Chat - very short, and hard to get context without copying

wizard kibitzes: Hello diane, good luck!
You kibitz: Hi from spring in Oz, enjoy!

wizard says: all my games tonight ended when opponents got 3 doubles during bear off
wizard says: wonder why it's always 3 and not 1 or 2 or 4 or 5, etc.

You kibitz: my thoughts are:
You kibitz: 1 - I bet it isnt *always* 3, save matches, analyse matches - including the ones you won, and check...i bet your human gut feeling is wrong - at least over a statistically significant number of matches
You kibitz: 2 - dont let it get to a race where 3 doubles will win it

wizard says: I'm just wondering why so many doubles at the end of every game, whether I get them or opponent gets them.
wizard says: ok, not every game, but maybe 75%.
wizard says: hmmmm, that's the second time with the 44

then he resigns B...and invites me again.

Oddly, I wasn't really interested in playing another whine fest, so I didn't accept..

Message from RepBotNG: wizard has complained about you.
wizard says: I complained because you won't give me a match.
wizard says: bye

message wizard and I complained about you, gagged you, and posted about you for all fibsters to see, because you are a dice whining bore...

Here are the rolls...as many of them as we managed to get before he resigned rather than even try to play..
W     Me
        53:
  41: 56:
  43: 52:
  65: 65:
  14: 44:
  21: 46:
  21: 21:
  61: 35:
  65: 26:
  43: 44:
  25: 42:
  26: 34:
  53: 44:
  22: 16:
  53: 43:
  63: 64:

Apart from my slight predisposition to roll 44, anyone else feel this is a misrepresentative sequence of 31 rolls of two dice? 

the "unusual" 6-2, 2-6, 6-2, 2-6, 4-4, 4-4, 4-4, 4-4 is as likely as the "normal" 5-3, 4-2, 5-4. 3-1, 2-2, 3-1, 5-6, 2-3.

Sounds like a British-funded revolutionary amendment project in probability theory to me. Would you elaborate that how can the probabilities of the first and second sequences are 'equal'?

PL

Yes. The chance of any roll occurring with two six-sided dice is of course 1/36, i.e any face on one dice with any face on the other. The chance of rolling 6-6 is the same as the chance or rolling 6-2, or indeed 2-6. 6-2 and 2-6 appear the same when we are playing backgammon, because there is no need to differentiate between them, but in terms of a probable event they are seperate and distinct. Thus the probability of either of the two sequences that I offered as examples is 1/36 to the power 8.
One appears "normal", the other abnormal, but they are neither. They are just two short strings in the endlessly wonderful dice journey. Enjoy.

So you wouldn't complain if you rolled 6-2 forever since 6-2 6-2 6-2 ... is just as likely as all other sequences?

No. Compaining about something over which one has no control is a waste of time. However I did once take part in a chouette where the captain of the team switched the box's dice for a pair that had all fives on one die and all twos on the other. It took him eleven rolls to catch on, in the end because by this time the team were rolling on the floor and clutching their sides in hysterical laughter. He didn't see the funny side of it and left in a huff.

Good story. What I was trying to point out was that according to your philosophy they shouldn't have been laughing at all: Since the sequence of combinations of 5s and 2s which the box rolled with the rigged dice has the same probability of occuring as any other (with normal dice), the box would have been justified in not getting suspicious about his rolls.

My point is that saying that "every sequence occurs with the same probability" is not a useful argument to counter accusations of the dice being rigged. A long sequence of 5s and 2s is indeed indicative (although not a 100% proof) of something being wrong with the dice. It does constitute a "misrepresentative" set of rolls.

If you rolled 6-2 20 times in a row on fibs and then had to bet money on whether there is something wrong with the dice generator, what would you do? (I would probably leave in a huff...)

However, you have grasped what some do not, that an endlessly repeating sequence is as likely as any other. It is no more or less common, merely more visible.

It would be equally wrong to say that winning a 60-pips-behind race by rolling 66 66 66 on the last three rolls cannot be called a supremely lucky event because that sequence has the same probability as all other 3 roll sequences. One *might* counter with saying that rolling the 66 triplet evened out the total luck in the match (provided of course that that is true for the match in question).

Depending on the perspective taken, the roller of the 66 triplet might be called lucky or not. Still, rolling 66 66 66 at *some* point in the game isn't the same as rolling it on the last three rolls when all that saves you from losing is just that sequence.

Likewise, when people complain about supposedly rigged dice on fibs, they aren't referring to the fact that the sequence they rolled occurred at *some* point (in which case "all sequences have equal prob." would indeed be a correct counter), but that it occured precisely when it was useful to their opponent.

The fact that luck evens out to zero over time does not mean that there is no such thing as a lucky or unlucky roll. That is platitude. But there is a similar fallacy behind countering "the dice are rigged" with "all sequences of rolls have equal probability".

It wasn't my intention to counter allegations of dice rigging. That is a different matter entirely.

My point applies in equal manner to the statement that there are no "unusual" or "surprising" sequences.