none of them are quite what I want works not all work out. I'm not just looking for the largest sum I'm looking for the largest sum <= (n-1) such that the values are either the same one or right next to each other. for example for (7-1) = (1+1+2+2) = 3 + 3, not the 1 and 6 your function finds. and the index of those helps to find out what diagonal to check. I'll try to make one I'm not asking for you to have to support my idea.

[CODE]for(x=2,#mersenne1,if(mersenne1[x]-triangle[prectriM((mersenne1[x]-1)/2)]-triangle[prectriM((mersenne1[x]-1)/2)+1]-1>=0,print(prectriM((mersenne1[x]-1)/2)","prectriM((mersenne1[x]-1)/2)+1","mersenne1[x]-triangle[prectriM((mersenne1[x]-1)/2)]-triangle[prectriM((mersenne1[x]-1)/2)+1]-1","mersenne1[x]),print(prectriM((mersenne1[x]-1)/2)","mersenne1[x]-2*triangle[prectriM((mersenne1[x]-1)/2)]-1","mersenne1[x])))

1,0,3
1,2,0,5
2,0,7
3,0,13
3,4,0,17
3,4,2,19
5,0,31
7,4,61
8,9,7,89
9,10,6,107
10,11,5,127
22,14,521
24,6,607
35,18,1279
46,40,2203
47,24,2281
56,24,3217
64,65,27,4253
66,0,4423
97,98,84,9689
99,40,9941
105,82,11213
140,141,55,19937
146,147,91,21701
151,152,104,23209
210,186,44497
293,100,86243
331,332,278,110503
362,363,279,132049
464,330,216091
869,808,756839
926,927,103,859433
1121,24,1257787
1181,1182,1144,1398269
1724,1725,595,2976221
1737,1738,732,3021377
2640,352,6972593
3669,1686,13466917
4581,4582,1286,20996011[/CODE]

for some reason this doesn't match up unless I did the math wrong the first time. okay I did , I must have lost sight of the proper ones to subtract in the b file for A000217.

I see why 23209 and 3217 line up :

56/2 leaves 28 below the center

since 104 > 152/2 because the second one hasn't been done for 152 then we get 104- (152/2) = 28 so it to in this case is 28 below the center and lines up with 3217.

the hard part for me is turning this into a Cartesian plane then I could just subtract coordinates and figure the lineups out. I know what I'd have to do but it's complicated. because i have to use if the 2 first values in the 4 value ones in my list to tell which quadrant it's in and do all the subtraction I've done to show that the y coordinate is the same for 3217((-4,-28) I believe and 23209((-75,-28) if I've done the math correctly) then it's as simple as checking check if the x or y values are the same ( if not it then could check the difference in both and check if it's a 1 to 1 ratio).

After 200 posts in this thread... where are you going with all of this?? Several people have asked you several times, and you just go back to talking about "stuff" instead of answering the question. Or am I just missing something?

[QUOTE=science_man_88;251118]the hard part for me is turning this into a Cartesian plane then I could just subtract coordinates and figure the lineups out. I know what I'd have to do but it's complicated. because i have to use if the 2 first values in the 4 value ones in my list to tell which quadrant it's in and do all the subtraction I've done to show that the y coordinate is the same for 3217((-4,-28) I believe and 23209((-75,-28) if I've done the math correctly) then it's as simple as checking check if the x or y values are the same ( if not it then could check the difference in both and check if it's a 1 to 1 ratio).[/QUOTE]

okay kinda badly phrased. Basically I'd use the first number as the origin and see if I can plot the lines of slope 1/1 or horizontal or vertical lines between them.The problem I think I'd have is figuring out the position each one would have on the Cartesian plane when saying each box is another coordinate.

[QUOTE=KingKurly;251123]After 200 posts in this thread... where are you going with all of this?? Several people have asked you several times, and you just go back to talking about "stuff" instead of answering the question. Or am I just missing something?[/QUOTE]

my most recent idea starts on post 169.

basically I think I'm having trouble going from a number spiral ( where the idea was detected) to making the placement in the number spiral into Cartesian coordinates using the center as the origin (because lines seem easier on a Cartesian plane).

going around the number spiral if the 1 is at (0,0) then 2 is at (1,0) , 3 is at (1,1) since these share a x value then they are on the same line in this case the line x=1. the 5 is at position (-1,1) since the y coordinate is the same for the 3 and 5 they are in a line with each other y=1.that's all the transforming that's needed assuming you have the number spiral but with a position unknown in one it's hard to find it for the other. I'm just having trouble turning what I know about finding it in one into a way to find the coordinates with a list of coordinates it should be simple addition and subtraction to prove my idea.

okay I might be able to do this I have a way to get the coordinates of each corner:

[CODE]x=0;y=0;mode=1;for(z=1,10,if(mode==1,x=x+z;print(x","y);y=y+z;print(x","y);mode=0,x=x-z;print(x","y);y=y-z;print(x","y);mode=1))[/CODE]

[CODE]for(i=1,#mersenne1,
a=mersenne1[i]-1;
x=0;
y=0;
corner=0;
mode=1;
for(z=1,5000,
if(mode==1&&a>=z,
x=x+z;
a=a-z;
corner=corner+1;
if(a>=z,
y=y+z;
a=a-z;
corner=corner+1;
mode=0
)
,if(mode==0&&a>=z,
x=x-z;
a=a-z;
corner=corner+1;
if(a>=z,
y=y-z;
a=a-z;
corner=corner+1;
mode=1
)
)
)
);
if(corner%4==1,
print(x","y+a","a),
if(corner%4==2,
print(x-a","y","a),
if(corner%4==3,
print(x,","y","a),
if(corner%4==0,
print(x+a","y","a)
)
)
)
)
)[/CODE]

this is the best I've got it's not perfect though it makes errors 4 times in the first 26 I think. each time it's minus [B]a[/B] but unless i can tell what's making it mess up I can't fix it to be accurate for all of them.