rollup pool question(s)

Posted: Sat Aug 20, 2011 11:47 am

Just wondering if there's ever been any concerted effort to get like, averages or medians of the numbers in the rollup pool. I've played off and on since 1991, but I never really documented my raw rollups so I don't have a lot of data to figger with. It's always seemed to me that there's a disproportionate mess of low will rollups, is this just my feeling or is there any evidence one way or the other?

Posted: Sat Aug 20, 2011 11:49 am

(And should I have posted this to the rollup forum? It's more a question about the data gathering, but still...?)

Posted: Sat Aug 20, 2011 4:37 pm

saucyjack wrote:

(And should I have posted this to the rollup forum? It's more a question about the data gathering, but still...?)

Yes, the rollup forum would be better.

There may be varying pools - arenas, tourneys, tourney arenas, etc.
However, we combine ours into one "file".
We have kept a listing of our last 100+.
13 14 9 6 11 7 10
9 10 9 9 9 9 15
13 4 17 8 4 12 12
10 10 13 6 7 10 14
13 14 10 10 5 12 6
10 14 6 14 10 6 10
9 10 9 11 8 9 16
5 7 13 8 12 19 6
3 14 11 14 9 9 10
10 11 9 11 12 10 7
12 7 14 8 10 12 7
9 9 8 12 9 10 13
7 8 12 6 12 15 10
6 14 9 8 10 10 13
9 11 12 10 6 7 15
12 3 12 12 10 5 16
16 6 12 9 4 12 11
4 14 10 9 18 10 5
7 15 15 3 10 12 8
18 11 5 9 4 18 5
5 10 10 13 10 10 12
14 11 3 14 10 9 9
7 14 7 10 12 16 4
5 13 10 13 4 10 15
13 12 6 5 9 17 8
5 14 10 15 13 4 9
3 14 17 11 10 7 8
12 7 10 4 16 11 10
13 9 16 6 10 3 13
10 10 17 4 16 9 4
3 9 13 14 13 10 9
15 9 13 10 11 3 9
9 10 19 4 9 9 10
10 16 6 11 14 4 9
4 13 11 13 11 8 10
10 17 12 15 3 9 4
9 12 12 10 3 15 9
6 11 14 12 8 9 10
5 11 8 13 16 14 3
3 17 12 9 8 6 15
14 9 12 4 19 8 4
8 5 11 6 9 10 21
12 7 16 9 7 5 14
9 11 9 9 5 15 12
3 17 12 9 8 6 15
9 10 12 8 8 13 10
4 9 12 10 13 9 13
9 12 12 9 6 12 10
4 17 10 4 11 10 14
11 11 10 12 9 8 9
5 10 12 4 19 10 10
10 12 12 10 10 13 3
8 9 6 13 10 11 13
9 4 11 8 13 15 10
15 6 12 7 6 13 11
4 10 14 19 8 5 10
10 12 4 5 9 17 13
7 3 17 9 11 10 13
6 12 14 10 10 5 13
5 3 9 17 9 12 15
12 13 10 9 6 10 10
15 9 18 4 9 10 5
9 16 10 4 13 12 6
11 11 9 16 3 15 5
3 17 9 16 8 8 9
8 9 17 6 15 9 6
9 8 9 14 10 11 9
3 7 21 3 12 16 8
9 8 18 12 10 8 5
7 10 17 5 5 17 9
9 15 10 17 17 3 11
9 12 13 7 9 10 10
11 17 10 10 13 5 4
13 14 8 12 5 9 9
14 3 11 16 4 13 9
10 5 15 3 14 11 12
10 15 8 7 9 9 12
13 8 13 12 3 12 9
13 12 9 9 13 10 4
14 11 6 10 4 15 10
5 10 8 8 11 12 16
13 10 7 5 16 9 10
9 10 11 13 10 4 13
17 9 14 12 5 9 4
12 9 11 10 7 9 12
10 9 15 3 9 9 13
9 10 9 9 9 9 15
9 8 11 12 13 8 9
11 8 13 16 3 10 9
5 12 10 7 16 8 12
4 10 12 13 14 3 14
8 10 16 13 8 5 10
9 11 15 9 13 9 4
11 6 15 15 5 8 10
11 19 11 9 6 9 5
13 10 7 17 4 9 10
8 14 12 8 10 13 5
14 15 11 4 10 11 5
9 7 13 10 13 4 14
8 10 14 7 8 13 10
5 13 16 13 4 13 6
11 12 12 8 3 10 13
10 9 11 10 10 10 10
10 6 12 10 9 13 10
4 13 11 13 11 13 5
5 9 12 5 14 16 9
8 11 13 10 11 13 4
12 11 13 5 12 9 8
11 12 13 13 10 4 7
10 11 19 11 3 12 4
12 5 8 12 14 5 14
12 9 8 13 12 4 12
10 10 8 4 12 5 21
14 10 8 10 10 9 9
5 10 14 8 7 13 13
9 9 13 14 10 9 6
8 10 14 12 11 8 7
15 6 10 4 16 16 3
8 10 14 12 11 8 7
3 9 17 10 3 16 12
11 6 16 12 5 8 12
9 10 14 5 16 9 13
5 15 9 17 5 13 6
10 17 7 10 4 13 9
13 8 14 8 8 9 10
3 10 17 14 6 11 9
6 6 17 7 14 10 10
14 6 10 12 12 3 13
13 6 12 18 5 11 5
4 4 18 11 12 16 5
9 9 11 11 12 12 6
13 14 19 6 3 7 8
18 5 14 5 11 12 5
10 11 9 11 12 10 7
10 10 16 12 7 10 5
5 11 8 13 16 14 3
9 9 9 10 9 12 12
4 10 10 9 16 9 12
12 6 16 10 16 7 3
Use it as you see fit.

Posted: Sat Aug 20, 2011 5:09 pm

This is my most recent list if someone wants to crunch numbers. These are replacement RU's, I track my team setups seperately.

17 10 5 9 9 13 7
13 5 9 12 9 9 13
6 5 9 10 14 9 17
14 4 10 10 10 16 6
10 10 10 10 3 11 16
13 4 10 10 13 9 11
13 12 10 9 12 10 4
4 10 11 11 5 13 16
9 11 11 6 13 8 12
7 13 11 3 13 13 10
5 12 11 13 8 4 17
7 9 11 8 13 9 13
14 4 13 9 7 12 11
7 13 13 10 13 4 10
10 8 14 9 9 7 13
11 4 14 11 5 9 16
10 4 14 4 14 8 16
13 6 14 9 9 14 5
6 13 14 11 9 4 13
10 16 14 4 12 9 5
9 8 15 12 4 13 9
7 9 15 8 17 5 9
9 11 16 10 7 5 12
11 6 16 10 4 14 9
14 9 16 5 5 9 12
4 8 17 10 8 10 13
11 6 17 12 4 7 13
5 8 21 10 7 9 10
10 14 15 8 10 6 7
15 12 11 12 9 5 6
13 9 13 9 12 10 4

Posted: Sat Aug 20, 2011 5:50 pm

Thanks for the info, I will definitely put it to use!

ArchMaster Poster Joined: Apr 15, 2004 Posts: 2958

Pretty neat info guys.

I imported your data into a spreadsheet. I had to eliminate a few of Consortium's that didn't add up to 70:

9 10 9 11 8 9 16 72
3 9 13 14 13 10 9 71
9 15 10 17 17 3 11 82
10 9 15 3 9 9 13 68
11 12 12 8 3 10 13 69
9 10 14 5 16 9 13 76

But that still left us with 164 entries.

Average stats: 9.34-9.98-11.97-9.68-9.52-9.90-9.62

It seems they accounted for the fact that you couldn't add to SZ during their creation process, since SZ is 2.3 higher than the average of the other stats.

Of the 164:
12 that were capable of 21 WT.
14 that were capable of 21 WL.
16 that were capable of 21 DF.
17 that were capable of 17 WT/17 WL.
4 that were capable of 17 WT/17 WL/17 DF.

There are some overlap in these categories, so there aren't really 63 high potential roll-ups in the bunch, only 55 individual ones (which is still pretty high). Also, I did that part by hand, so I could have missed some. There are also some duplicates, but I didn't weed them out.

I would estimate maybe 18 of these make solid tournament warriors, with a number of others that are close. This is purely subjective, but even so its not a bad percentage.

Makes me wish I've been keeping track of my Dark Arena replacements!

Posted: Sun Aug 21, 2011 1:31 am

I love the analysis OAB!

The tourney caliber warriors must have come from Consortium's pool Very Happy

Posted: Sun Aug 21, 2011 1:44 am

Madwand wrote:

I love the analysis OAB!

The tourney caliber warriors must have come from Consortium's pool Very Happy

Maybe the 82 point guy? Wink

OAB, thanks for the heads-up; apparently I have a couple of typos/errors. I will delete them.

Posted: Sun Aug 21, 2011 6:34 am

One Armed Bandit wrote:

Pretty neat info guys.

I imported your data into a spreadsheet. I had to eliminate a few of Consortium's that didn't add up to 70:

9 10 9 11 8 9 16 72
3 9 13 14 13 10 9 71
9 15 10 17 17 3 11 82
10 9 15 3 9 9 13 68
11 12 12 8 3 10 13 69
9 10 14 5 16 9 13 76

Yeah, I saw that one and started adding adding 14 points to it, and went "Holy %@#$", 11-17-10-21-21-3-13, what a DM God..... wait a minute...

Posted: Sun Aug 21, 2011 9:07 am

_Buri_ wrote:

One Armed Bandit wrote:

Pretty neat info guys.

I imported your data into a spreadsheet. I had to eliminate a few of Consortium's that didn't add up to 70:

9 10 9 11 8 9 16 72
3 9 13 14 13 10 9 71
9 15 10 17 17 3 11 82
10 9 15 3 9 9 13 68
11 12 12 8 3 10 13 69
9 10 14 5 16 9 13 76

Yeah, I saw that one and started adding adding 14 points to it, and went "Holy %@#$", 11-17-10-21-21-3-13, what a DM God..... wait a minute...

I have corrected/deleted the "bad data".
The particular one in question resulted in a 11-15-10-17-17-3-11 WS. We do not know what the actual 70 point roll was.

Posted: Sun Aug 21, 2011 5:59 pm

That encouraged me to add a double-check column to my own spreadsheet, since I've already messed up on one set of numbers on these forums already. Adding in the rollups I do have documented that I've found so far (I think I have like...190 rollups or so counting the above), the averages are still so close to OAB's values it is not worth talking about yet.

Posted: Sun Aug 21, 2011 6:15 pm

That is very cool analysis. Thanks to those who contributed raw data or played with the numbers. I also like what it says about the replacement, they really do seem to be random - spread equally all over the board.

As for the one of the comments in the original post - does it seem like there are a lot of low will roll-ups? The #s seem to say, "No." I suspect the reason it seems that way is that mostpeople don't add any points to will unless the warrior can get to at least 15, maybe 17. It means there are a lot of 7 or 9 will warriors out there. Speed and con might have similar add patterns. Just a guess.

Unchartered Poster Joined: Aug 02, 2011 Posts: 18

They seem to be on a bell curve, if you ask me.

I suspect they roll 3 random numbers, maybe a 7-sided dice since the minimum value is 3 and the maximum value is 21. And perhaps on size, if they roll a value of 3 or 4, it is rerolled... since that number seems oddly disproportionate to the rest.

ST: Min = 3, Max = 18, Mean = 9.3, Median = 9, Std Deviation = 3.588
CN: Min = 3, Max = 19, Mean = 10, Median = 10, Std Deviation = 3.453
SZ: Min = 3, Max = 21, Mean = 11.9, Median = 12, Std Deviation = 3.448
WT: Min = 3, Max = 19, Mean = 9.7, Median = 10, Std Deviation = 3.403
WL: Min = 3, Max = 19, Mean = 9.5, Median = 10, Std Deviation = 3.729
SP: Min = 3, Max = 19, Mean = 9.9, Median = 10, Std Deviation = 3.456
DF: Min = 3, Max = 21, Mean = 9.7, Median = 10, Std Deviation = 3.823

Posted: Mon Aug 22, 2011 3:13 pm

Hmm. Now I wonder if each rollup is generated on the fly as a single unit, with each stat assigned from a diminishing pool of 70, or if a series of random numbers rolled on a given bell curve were generated and distributed until each individual unit hit 70, then gets called a 'rollup.' Just from anecdotal evidence in my own playing around with programming similar things, I wouldn't expect the results to be so consistent from such a small bunch of numbers if it was done one rollup at a time.

ArchMaster Poster Joined: Apr 15, 2004 Posts: 2958

I don't believe each roll-up is generated on the fly.

There were some duplicates in the data above and I've received duplicates from the Dark Arena before. I'm pretty sure the roll-ups were randomly generated when the game began, but now they just grab a replacement from that finite pool of roll-ups. The pool is rather large, so unless you run a lot of teams or play for decades, you might never notice.