Questions

[BuzzBuzz]

How are the points calculated that are awarded for solving a cryptogram?

[pootie49]

Please see this post for the answer to your question

http://www.cryptograms.org/forum/sho...hlight=scoring

basically an easy puzzle is scored less and you are awarded more points for avg and then difficult.

time also makes a difference if you solve faster you get higher points, Maradnu's calculations are
actually spot on. Based on times of twenty seconds and over.
The scores for solving under twenty are a bit higher.

[BuzzBuzz]

Thanks for the reply Pootie. If you tell me where I can find the actual formulas used, I would much appreciate it.

[pootie49]

You would have to send an email or private message to Admin to find out the actual
formulas although I am curious as to why?

[Lily H]

Why are you asking all these questions about scoring BuzzBuzz when you haven't even attempted to play a single game? Are you trying to create a rival cryptogram site or maybe you're writing a cheating program so you can get fast solve times?

[BuzzBuzz]

Hi Pootie and Lily:

Thanks for your comments.

I am a mathematician by training, and I am curious about some odd things I see related to the bell shaped curve that accompanies the time (in seconds) it takes for one to solve a cryptogram.

There are five numbers shown at the bottom of each bell shaped curve, which I assume are related to the statistics of the collection of times taken by all the solvers for one cryptogram , or perhaps just including solvers who are competing. For discussion purposes, I name these five values from left to right as follows: Min, Low, Mid, Hi, Max.

I discovered that the ratios of Max:Mid and Hi:Mid are always the same: 5.0 and 3.0 respectively. But, the ratios of Mid:Min and Mid:Low change for each cryptogram. For the five cryptograms I have just looked at, the Mid:Min ratio range in value from 4.67 to 6.78, and the Mid:Min ratio range in value from 1.65 to 1.73.

Statistically, this makes no sense.

[pootie49]

um, a whawhat?

Well you chose the right tactic to shut me up...you used math.
You might need to send a pm to admin and he can explain it or one of our many
extremely bright players will have a better answer.
Heck I have no idea. I simply am not able to grasp mathematical concepts.
All I do know how to do is solve and hope I have a decent time and maybe now and
then set a record.
erm hold on a tick....five cryptograms? Out of tens of thousands.
Is that a fair assessment?
I thought making a conclusion or having a theory required considerably more
data than that?

[pajarito7]

BuzzBuzz,
The Bell-shaped curve graphics to the contrary, the distributions (particularly for puzzles that have not been solved enough times) are not Gaussian. Admin is using the median for his average.

[BuzzBuzz]

Originally Posted by pajarito7

BuzzBuzz,
The Bell-shaped curve graphics to the contrary, the distributions (particularly for puzzles that have not been solved enough times) are not Gaussian. Admin is using the median for his average.

Hi Parajito:

Thank you for your help. I understand that the distibution cannot be Gaussian since negative solution times are not possible. I had guessed that the center of the bell shape curve shown was likely calculated as a median rather than a mean since that would be a more useful statistic. I appreicate your confirming this for me.

I would also like to understand why the Hi and Max values are calcualted respectively as exactly three times the median and five times the median. I am also curious about how the Min and Low values are calculated. Can you help me with that?

[BuzzBuzz]

Originally Posted by pootie49

um, a whawhat?

Well you chose the right tactic to shut me up...you used math.
You might need to send a pm to admin and he can explain it or one of our many
extremely bright players will have a better answer.

Hi Pootie:

I am sorry my math shut you up. That was not my intention. It was to get a response like the one pajarito sent. Thanks for your suggestion about sending a pm to admin. I will do that if this thread fails to give me that answers I seek.

You are right about five examples generally not being enough for a useful statistical analysis. However, for the particular results I got (unchanging values for all five examples) for two of the ratios, five examples are sufficient.