This kind of FindMinimum is something I have often done. Now it's gone crazy.

Look carefully at the search descriptions for the two versions of FindMinimum. In the first version, u0 is gray, as though it is unknown.

In the second version, I have moved the u search description to the end. So now the x20 is grayed, although it wasn't grayed before! So somehow, the first search description is screwed up, no matter what it is.

At the top, you see that u0 and x20 have perfectly reasonable values.

If I start either one of the FindMinimum's, I get very wrong behaviors ("err" is instrumented to tell me what is happening.)

I've stared at this a million times, and I don't see the problem. Any suggestions?

I have attached a part of the notebook. The whole notebook is long, and works off a very long data file. I have left the screen shot attached, since it shows the colors I am talking about. The code sample does not show these.

In[411]:= {x20, u0}

Out[411]= {-55, 10.}

count = 0;
result = FindMinimum[
  Hold@err[
    bez5[#, {llc[u], {x2, y2}, {x3, y3}, {x4, y4}, ulc[v]}, {w10, w2, 
       w3, w4, w50}] &,
    lcData,
    Table[1, Length@lcData],
    {u, {llc[u], {x2, y2}, {x3, y3}, {x4, y4}, ulc[v]}, 
     v, {w10, w2, w3, w4, w50}}
    ],
  {u, u0, u0 + .1},
  {x2, x20, x20 + 1}, {y2, y20, y20 + 1}, {w2, w20, w20 + .1},
  {x3, x30, x30 + 1}, {y3, y30, y30 + 1}, {w3, w30, w30 + .1},
  {x4, x40, x40 + 1}, {y4, y40, y40 + 1}, {w4, w40, w40 + .1},
  {v, v0, v0 + .1}
  ]

count = 0;
result = FindMinimum[
  Hold@err[
    bez5[#, {llc[u], {x2, y2}, {x3, y3}, {x4, y4}, ulc[v]}, {w10, w2, 
       w3, w4, w50}] &,
    lcData,
    Table[1, Length@lcData],
    {u, {llc[u], {x2, y2}, {x3, y3}, {x4, y4}, ulc[v]}, 
     v, {w10, w2, w3, w4, w50}}
    ],
  {x2, x20, x20 + 1}, {y2, y20, y20 + 1}, {w2, w20, w20 + .1},
  {x3, x30, x30 + 1}, {y3, y30, y30 + 1}, {w3, w30, w30 + .1},
  {x4, x40, x40 + 1}, {y4, y40, y40 + 1}, {w4, w40, w40 + .1},
  {u, u0, u0 + .1},
  {v, v0, v0 + .1}
  ]