Tervic2009-03-11 14:05:12
I want to do a statistical analysis of forging plate, mostly 'cause I'm really really bored right now. If anyone is willing, it's a fairly trivial matter to get ZMUD to output to a text file using the #file and #wri commands if you already have a forging script.
#file 1 forgingdata
will create a new textfile "forgingdata.txt" if it doesn't exist yet.
#wri 1 "blah blah blah"
will write the line "blah blah blah" to the file identified as #1, in this case forgingdata.txt.
The file format I'm using is a single value per line, with the first line being the type of forging data that has been catalogued (this first line is unnecessary, I can add that myself later).
Example:
ArmourBlunt
75
81
71
80
68
84
73
71
83
70
so yeah, if people are willing to upload textfiles with their data, I would really really appreciate it.
EDIT: Oh also, I'm looking for a way to analyze the distribution of the points between cutting and blunt, so that I can develop a model for predicting the % chance to get a xx/yy plate. Halp?
==================================================================
Most recent information will be posted here as I forge.
Last update: 3/12 11:55 PM
--I found a minor bug in my program where the blunt calculations were including all the data from the cutting calculations too... this has been fixed.
Cutting:
Total values analyzed: 1995
Average: 81.10175438596491
Minimum value: 57
Maximum value: 106
Standard Deviation: 7.397633702124045
Blunt:
Total values analyzed: 1994
Average: 80.86208625877633
Minimum value: 58
Maximum value: 106
Standard Deviation: 7.340994992712547
#file 1 forgingdata
will create a new textfile "forgingdata.txt" if it doesn't exist yet.
#wri 1 "blah blah blah"
will write the line "blah blah blah" to the file identified as #1, in this case forgingdata.txt.
The file format I'm using is a single value per line, with the first line being the type of forging data that has been catalogued (this first line is unnecessary, I can add that myself later).
Example:
ArmourBlunt
75
81
71
80
68
84
73
71
83
70
so yeah, if people are willing to upload textfiles with their data, I would really really appreciate it.
EDIT: Oh also, I'm looking for a way to analyze the distribution of the points between cutting and blunt, so that I can develop a model for predicting the % chance to get a xx/yy plate. Halp?
==================================================================
Most recent information will be posted here as I forge.
Last update: 3/12 11:55 PM
--I found a minor bug in my program where the blunt calculations were including all the data from the cutting calculations too... this has been fixed.
Cutting:
Total values analyzed: 1995
Average: 81.10175438596491
Minimum value: 57
Maximum value: 106
Standard Deviation: 7.397633702124045
Blunt:
Total values analyzed: 1994
Average: 80.86208625877633
Minimum value: 58
Maximum value: 106
Standard Deviation: 7.340994992712547
Casilu2009-03-11 16:08:54
QUOTE (Tervic @ Mar 11 2009, 07:05 AM) <{POST_SNAPBACK}>
I want to do a statistical analysis of forging plate, mostly 'cause I'm really really bored right now. If anyone is willing, it's a fairly trivial matter to get ZMUD to output to a text file using the #file and #wri commands if you already have a forging script.
#file 1 forgingdata
will create a new textfile "forgingdata.txt" if it doesn't exist yet.
#wri 1 "blah blah blah"
will write the line "blah blah blah" to the file identified as #1, in this case forgingdata.txt.
The file format I'm using is a single value per line, with the first line being the type of forging data that has been catalogued (this first line is unnecessary, I can add that myself later).
Example:
ArmourBlunt
75
81
71
80
68
84
73
71
83
70
so yeah, if people are willing to upload textfiles with their data, I would really really appreciate it.
#file 1 forgingdata
will create a new textfile "forgingdata.txt" if it doesn't exist yet.
#wri 1 "blah blah blah"
will write the line "blah blah blah" to the file identified as #1, in this case forgingdata.txt.
The file format I'm using is a single value per line, with the first line being the type of forging data that has been catalogued (this first line is unnecessary, I can add that myself later).
Example:
ArmourBlunt
75
81
71
80
68
84
73
71
83
70
so yeah, if people are willing to upload textfiles with their data, I would really really appreciate it.
If there was a way for someone to write one for MUSH, I'd help. At the rate I'm going, you'll have infinite trials.
Isuka2009-03-11 16:17:05
QUOTE (casilu @ Mar 11 2009, 09:08 AM) <{POST_SNAPBACK}>
If there was a way for someone to write one for MUSH, I'd help. At the rate I'm going, you'll have infinite trials.
After I switched from cmud to mush, I never rewrote my statistical tracking functions, unfortunately.
Edit: That said, I'm going to write one right now, and store every forge that I get in a MySQL table... and probably then make a website that will display trends for items.
Tervic2009-03-11 20:08:37
QUOTE (Isuka @ Mar 11 2009, 09:17 AM) <{POST_SNAPBACK}>
After I switched from cmud to mush, I never rewrote my statistical tracking functions, unfortunately.
Edit: That said, I'm going to write one right now, and store every forge that I get in a MySQL table... and probably then make a website that will display trends for items.
Edit: That said, I'm going to write one right now, and store every forge that I get in a MySQL table... and probably then make a website that will display trends for items.
Hopefully we can use this information to conclusively prove that forging sucks beyond all belief.
Casilu2009-03-11 20:31:59
QUOTE (Tervic @ Mar 11 2009, 01:08 PM) <{POST_SNAPBACK}>
Hopefully we can use this information to conclusively prove that forging sucks beyond all belief.
Well, if someone could get me a script for MUSH to save all these forges that I'm doing...
Tervic2009-03-12 00:07:50
From the data I've collected so far, it looks like the minimum you'll ever get on platemail is 55, the max is 110, the average is 81 with a standard deviation of 7.5.
This means that for a single stat to be 90 or above, you need to be 9/7 or 1.28 standard deviations over the mean, meaning that about 10% of armours will have a stat of 90 or above. I'll have to go back and look at the raw data, but this seems to be about right from what I've observed. If one treats the armour values as independent variables, there is a 1% chance for 90/90 or better, but since they're not independent, this is a flawed assumption. Anyways, we can see from my 1300 analyzed values that I haven't gotten a 90/90 yet (since that's when I'm going to stop forging). If each forging has a 1% chance, then I should have had a 1 - 2.11715657 × 10-6 chance to get the armour I desire. That's really freaking close to 100%, so it's clearly less than 1% to get a 90/90 armour.
This means that for a single stat to be 90 or above, you need to be 9/7 or 1.28 standard deviations over the mean, meaning that about 10% of armours will have a stat of 90 or above. I'll have to go back and look at the raw data, but this seems to be about right from what I've observed. If one treats the armour values as independent variables, there is a 1% chance for 90/90 or better, but since they're not independent, this is a flawed assumption. Anyways, we can see from my 1300 analyzed values that I haven't gotten a 90/90 yet (since that's when I'm going to stop forging). If each forging has a 1% chance, then I should have had a 1 - 2.11715657 × 10-6 chance to get the armour I desire. That's really freaking close to 100%, so it's clearly less than 1% to get a 90/90 armour.
Tervic2009-03-12 00:09:01
From the data I've collected so far, it looks like the minimum you'll ever get on platemail is 55, the max is 110, the average is 81 with a standard deviation of 7.5.
This means that for a single stat to be 90 or above, you need to be 9/7 or 1.28 standard deviations over the mean, meaning that about 10% of armours will have a stat of 90 or above. I'll have to go back and look at the raw data, but this seems to be about right from what I've observed. If one treats the armour values as independent variables, there is a 1% chance for 90/90 or better, but since they're not independent, this is a flawed assumption. Anyways, we can see from my 1300 analyzed values that I haven't gotten a 90/90 yet (since that's when I'm going to stop forging). If each forging has a 1% chance, then I should have had a 1 - 2.11715657 × 10-6 chance to get the armour I desire. That's really freaking close to 100%, so it's clearly less than 1% to get a 90/90 armour.
This means that for a single stat to be 90 or above, you need to be 9/7 or 1.28 standard deviations over the mean, meaning that about 10% of armours will have a stat of 90 or above. I'll have to go back and look at the raw data, but this seems to be about right from what I've observed. If one treats the armour values as independent variables, there is a 1% chance for 90/90 or better, but since they're not independent, this is a flawed assumption. Anyways, we can see from my 1300 analyzed values that I haven't gotten a 90/90 yet (since that's when I'm going to stop forging). If each forging has a 1% chance, then I should have had a 1 - 2.11715657 × 10-6 chance to get the armour I desire. That's really freaking close to 100%, so it's clearly less than 1% to get a 90/90 armour.
