numbers game

[Matt Coates]

im not sure this has been covered, but does anybody know of a selection where it is possible to get every number from 100 to 1000, would be interesting to know

[Joseph Bolas]

im not sure this has been covered, but does anybody know of a selection where it is possible to get every number from 100 to 1000, would be interesting to know

Not here, but in the mailing list this topic did appear. To quote Ray Folwell (who devised a program to work this out).

"According to my calculations 10,858,746 (91.2%) of the 11,905,457 possible numbers games can be solved.

As Howard pointed out, the most difficult selection is 1,1,2,2,3,3 which cannot produce anything larger than 81. I was surprised to find that there are 1,226 selections which can each be solved for any target between 101 and 999.

The most difficult target is 947 which can only(?) be reached from 9,017 selections. The easiest ones are 102,104 and 108 which can each be reached from 13,240 selections, just 3 short of the total number of selections."

EDIT: I don't know if this has answered your question, if not, you would need to speak to Ray, he will definitely know.

[Joseph Bolas]

After watching Friday's game and seeing how easy the numbers where, I was wondering if anyone knew the following.

According to Ray Folwell, 91.2% of the 11,905,457 possible numbers games that can be generated, can be solved. Has there though, been an episode, where all 3 number games that have been generated (in the 45-minute episodes), have not been solveable?

With the 30-minute format, there was only 2 number games so there's probably an occasion where this has happened.

[Michael Wallace]

Well if 8.8% of number games are not solvable, then that means the probability of 3 not solvable rounds in one show (assuming first that the contestants pick any one of the possible games uniformly, which is unlikely) is 0.000681472, so you'd expect one such show every 1468 episodes. In the old days, that reduces to once very 130 shows.

Of course, this isn't very helpful, since contestants can choose any one of five selections (and presumably are biased towards the 'easiest' 1 large), so for real number crunching you'd need to know the distribution of selections (which might be interesting anyway), and on top of that, the proportion of solvable games for each selection type.

By which stage it's probably easier just to go through the archives to find an 'unsolvable' game instead

[Charlie Reams]

According to Ray Folwell, 91.2% of the 11,905,457 possible numbers games that can be generated, can be solved.

That's the number of games that can be solved exactly. The number where it's possible to get within 10 is much higher, and I doubt there's ever been a game where both games were completely impossible to score on. Once the wiki project to tidy up all the recaps is complete, I'll be able to give some reasonably wide-based empirical data for this.

[Gavin Chipper]

Not here, but in the mailing list this topic did appear. To quote Ray Folwell (who devised a program to work this out).

"According to my calculations 10,858,746 (91.2%) of the 11,905,457 possible numbers games can be solved.

As Howard pointed out, the most difficult selection is 1,1,2,2,3,3 which cannot produce anything larger than 81. I was surprised to find that there are 1,226 selections which can each be solved for any target between 101 and 999.

The most difficult target is 947 which can only(?) be reached from 9,017 selections. The easiest ones are 102,104 and 108 which can each be reached from 13,240 selections, just 3 short of the total number of selections."

EDIT: I don't know if this has answered your question, if not, you would need to speak to Ray, he will definitely know.

Does all of this work on the basis that 101 is the minimum? I think there have been games with 100 as the target.

[Joseph Bolas]

I doubt there's ever been a game where both games were completely impossible to score on.

I agree with that. What I am saying is that, is there an episode where you could only get a maximum of 14/20 or 21/30 points on the numbers (therefore not perfectly solving any of the numbers)?

EDIT:

Does all of this work on the basis that 101 is the minimum? I think there have been games with 100 as the target.

Yes, these figures do focus on the minimum number being 101. There probably has been games in the past where where the target has been 100. If you were to go as far back as Calendar Countdown, when they used ERNIE (I think thats what they called it), because of the '0' on the first reel, you could have targets under 100.

[Jennifer Turner]

If you were to go as far back as Calendar Countdown, when they used ERNIE (I think thats what they called it), because of the '0' on the first reel, you could have targets under 100.

I think it was "Harry". On one of those retrospective programmes (probably "The 2000th Edition") there was a clip of Richard Whiteley conducting a numbers game on Calendar Countdown and he says "Let's see what Harry's got for us". I suppose it's possible (but unlikely in that context) that Harry could have been the contestant's name. I would assume that Harry, if that was the name, was named along similar lines to CECIL, after some Yorkshire TV bigwig.

[Joseph Bolas]

If you were to go as far back as Calendar Countdown, when they used ERNIE (I think thats what they called it), because of the '0' on the first reel, you could have targets under 100.

I think it was "Harry". On one of those retrospective programmes (probably "The 2000th Edition") there was a clip of Richard Whiteley conducting a numbers game on Calendar Countdown and he says "Let's see what Harry's got for us". I suppose it's possible (but unlikely in that context) that Harry could have been the contestant's name. I would assume that Harry, if that was the name, was named along similar lines to CECIL, after some Yorkshire TV bigwig.

I have double-checked UK Gameshows and it does say Harry, so you were right :

On the earlier Calendar Countdown, the numbers were generated by a mechanical device with spinning drums called, for some reason, Harry. Harry's "hundreds" drum had a zero on it, so potentially anything from zero to 999 could have turned up.

I don't know where I got the name ERNIE from .

[Michael Wallace]

I don't know where I got the name ERNIE from .

ERNIE is the premium bonds number generator-thing, so I'd guess that?