Re: random ints on a large symmetric interval



On Mon, 22 Sep 2008 13:56:17 -0600, Ron Ford posted:

On Sat, 20 Sep 2008 15:38:42 -0600, Ron Ford posted:

One thing about producing symmetric integers like this is that the expected
value is zero. Anyways, I just wanted to throw this out there and see what
other eyes say. I've got almost all of the integers in the right place
now, and I'll need to fiddle with the intervals, but I think I've got all
all the parts declared right now to get a flat distribution on an interval
that is appropraitely declared i13.

Why is my expected value the maximum value of the smaller interval instead
of zero?

The intervals needed tweaking. I think I have a better version now:

[code elided]

I became interested in the process of getting this done and took a couple
screenshots as I was working on this. As I've had a good interaction with
gfortran people, I was thinking of it from a gfortran.pdf point of view.

In short, I think it needs more pictures. I think this one shows what an
incredible tool an IDE can be:

http://i36.tinypic.com/281y8np.jpg

We see good behavior in this program by virtue of the expected value being
about negative fifty thousand with a much larger trial size. That's pretty
close to zero and south of the border.

§
How does one invoke gfortran? The answer differs among persons. For me,
gfortran is on a stick that belongs in my left pocket during transit. It
could be my shadow in that it seems to follow me in interesting ways. I've
been down for two days because my bios were set wrong with respect to usb
devices. My sysadmin buddy who believes that linux is one of the unices
resolved this conflict as I cooked noodles, sauteed onions and garlic, and
sliced the spam. (Gray spam doesn't taste good but probably will sustain
you. I had the luxury of spitting it out)

This screenshot shows how to get the better part of gfortran:
http://i35.tinypic.com/2hfumgm.jpg

I've got the correct goocher commented out by silverfrost for when I have
to make a transition. The transition is complete when the goocher comment
is dereferenced onto the gfortran command line. Since the compile time was
"fresh," I could tell that the gfortran compile time was a tenth or less.


I've got to clean and vamoose. I'll put output I don't understand entirely
after the man who procephied of George Bush: H. L. Mencken.

--
To die for an idea; it is unquestionably noble. But how much nobler it
would be if men died for ideas that were true!
H. L. Mencken


F:\gfortran\source>gfortran -o pop -Wall freeformat65.f95

F:\gfortran\source>pop
imax 2= 100000 itot2= 5800001
interval2= 200001 imax3= 2900000
n= 8
clock= 392856272
seed= 392856272 392856309 392856346 392856383 392856420
392856457
392856494 392856531
1 0.32027084
2 0.97826737
3 0.81613988
4 0.93265474
first ten elements of m are -2993973 -2983444
-2973165 -2969731 -2966185
-2965202
-2962094 -2960391 -2954718
-2954452

middle ten elements of m are 24606 24764
25242 25282 25894
26176
26472 26611 26661
2670
4 28361
last ten elements of m are 2956947 2957125
2959302 2964377 2964469 2969397
2971846 2972435 2980298
2990450
inbounds = 10000
outbounds = 0
percent = 0.0000000 %
summa s is 57129888.
expected value s is 5146.0068
summa t is -5669819.0
expected value t is -566.98187
expected value m is 5146.0068

F:\gfortran\source>gfortran -o pop -Wall freeformat65_2.f95

F:\gfortran\source>pop
imax 2= 100000 itot2= 599800001
interval2= 200001 imax3= 299900000
n= 8
clock= 393293210
seed= 393293210 393293247 393293284 393293321 393293358
393293395
393293432 393293469
1 0.85002458
2 0.90149671
3 3.28981876E-02
4 0.99657613
5 0.62373912
6 0.59350926
7 0.27234364
8 0.38576776
first ten elements of m are -299952257 -299932839
-299867276 -299846131 -299825164
-299684548
-299631558 -299586402 -299582002
-299572535

middle ten elements of m are 1027719 1034455
1122756 1168688 1170192
1205908
1277206 1331485 1478138
154620
7 1731935
last ten elements of m are 299297665 299357591
299440467 299576302 299594248 299620852
299647812 299774317 299785210
299949267
inbounds = 10000
outbounds = 0
percent = 0.0000000 %
summa s is 1.38257367E+10
expected value s is 1383326.9
summa t is 7532900.0
expected value t is 753.28998
expected value m is 1383326.9

F:\gfortran\source>gfortran -o pop -Wall freeformat65_2.f95

F:\gfortran\source>pop
imax 2= 100000 itot2= 599800001
interval2= 200001 imax3= 299900000
n= 8
clock= 393343694
seed= 393343694 393343731 393343768 393343805 393343842
393343879
393343916 393343953
1 0.79422665
2 0.84918892
3 0.86512077
4 0.73312682
first ten elements of m are -299795732 -299782552
-299779615 -299700986 -299611716
-299531467
-299529348 -299523347 -299337737
-299284777

middle ten elements of m are 906616 938294
1070461 1249044 1317141
1360340
1618146 1629768 1661720
169409
9 1727858
last ten elements of m are 299405761 299418433
299480691 299480809 299514328 299539477
299595019 299615084 299837463
299851716
inbounds = 10000
outbounds = 0
percent = 0.0000000 %
summa s is -6.75681178E+09
expected value s is -676210.06
summa t is -5289079.0
expected value t is -528.90790
expected value m is -676210.06

F:\gfortran\source>
.