what is the model you want to fit the data against? for me it is too hairy to fiddle that from the sheet ...

just guessing: if it is towards Nash equilibrium then other methods should be known

just guessing: if it is towards Nash equilibrium then other methods should be known

At least that gives errors, using a=365.0 as a float returns 0,
may be someone coded "expr/Gamma(negativeInteger) = 0" ?

To reach your identity one can enforce the use of the reflection
formula:

  restart;

  p := product((a-n)/a, n=1..N);

    -(-1)^(N+1)*GAMMA(N+1-a)*(1/a)^(N+1)/GAMMA(1-a)*a

  G := x -> Pi/sin(Pi*x)/GAMMA(1-x);

  eval(p,GAMMA=G); 
  expand(%); 
  simplify(%) assuming N::nonnegint;

                             (-N)
                            a     GAMMA(a)
                            --------------
                             GAMMA(a - N)


Note that "eval(p,GAMMA=G); eval(%,a=365.0);" results in 0 !@#!#!

At least that gives errors, using a=365.0 as a float returns 0,
may be someone coded "expr/Gamma(negativeInteger) = 0" ?

To reach your identity one can enforce the use of the reflection
formula:

  restart;

  p := product((a-n)/a, n=1..N);

    -(-1)^(N+1)*GAMMA(N+1-a)*(1/a)^(N+1)/GAMMA(1-a)*a

  G := x -> Pi/sin(Pi*x)/GAMMA(1-x);

  eval(p,GAMMA=G); 
  expand(%); 
  simplify(%) assuming N::nonnegint;

                             (-N)
                            a     GAMMA(a)
                            --------------
                             GAMMA(a - N)


Note that "eval(p,GAMMA=G); eval(%,a=365.0);" results in 0 !@#!#!

Ok. But looking at x^2 + const (or similar) shows that this is a bit senseless as question, as it is too general - or needs care for the special case(s).

Ok. But looking at x^2 + const (or similar) shows that this is a bit senseless as question, as it is too general - or needs care for the special case(s).

  restart;
  Digits:=14;

  d:=proc(x)
    local q,a,b;
    q:=convert(evalf(x),rational);
    b:=1/denom(q);
  end proc;
 
  plot('d'(x),x=0.1..Pi, thickness=3, numpoints=100);

Find below my older stuff: essentially it modifies the the Compile command,
such that system call are written to batch files (which can be kept to check
what is going on), which then are called to execute the according commands.

I encourage you to write your needed modifications into such bat files, 
there will be 2 of them (compiling and linking) each time.

Modifying the Compile command is temporary (i.e. the are not stored and will
be forgotten after closing the Maple sheet - hence do not modify Maple, but
have to be executed (throug an *.ini file) after a restart).

And as far as I remember with some experience it also was possible to give
readable names to the files (currently they are machine generated by Maple).


PS: does someone know, how to set MAPLE_TEMPDIR to a user-defined value on
Win XP? I completely ignored to care for such settings after Win ME ...
Currently C:\Dokumente und Einstellungen\Axel Vogt\Lokale Einstellungen\Temp
is where I find the resulting stuff ...

PPS: The reason for me was, that Win ME did not accept such long inputs for
its command lines. And I think, this way one can put several compiled ftcs
into one re-usable DLL.

###########################################################################

  kernelopts(opaquemodules=false):
  
  Compiler:-Build:-CompileSourceFile:=proc(oFile::string, cFile::string)
  # modified by Axel Vogt
  local cmd, status, batchFile, fd;
  
  cmd := Compiler:-Build:-compileCommand(oFile,cFile);
  if cmd <> NULL then
    Compiler:-` Compiler`(3,sprintf("Compile Command: %s",cmd));
    #status := Compiler:-Build:-ssystem(cmd);
  
    # use a batch file instead sending command to system
    batchFile:=StringTools:-Substitute( oFile, ".obj", "_for_OBJ.BAT" );
    fd := fopen(batchFile, WRITE);
      fprintf(fd, "%s",cmd);
    fclose(fd);
    status := Compiler:-Build:-ssystem(batchFile);
  
    if status[1] <> 0 then
      error "shit: compiler exited with nonzero status %1: %2", op(status)
    end if;
  
    # remove the batch file
    try
      #FileTools:-Remove(batchFile) # <--- activate to delete the batch file
    catch :
      NULL
    end try;
  
    true
  else
    Compiler:-` Compiler`(3,"No Compile Command.");
    false
  end if
  end proc:
  
  Compiler:-Build:-LinkDSO:= proc(soFile::string, oFile::string, cFile::string)
  # modified by Axel Vogt
  local cmd, status,  batchFile, fd;
  
  cmd := Compiler:-Build:-linkCommand(soFile,oFile);
  #print(cmd);
  if cmd <> NULL then
    Compiler:-` Compiler`(3,sprintf("Link Command: %s",cmd));
    #status := Compiler:-Build:-ssystem(cmd);
  
    batchFile:=StringTools:-Substitute( oFile, ".obj", "_for_DLL.BAT" );
    fd := fopen(batchFile, WRITE);
      fprintf(fd, "%s",cmd);
    fclose(fd);
    status := Compiler:-Build:-ssystem(batchFile);
  
    if status[1] <> 0 then
      error "shit, linker exited with nonzero status %1: %2", op(status)
    end if;
  
    # remove the batch file
    try
      #FileTools:-Remove(batchFile) # <--- activate to delete the batch file
    catch :
      NULL
    end try;
  
  else
    Compiler:-` Compiler`(3,"No Link Command.")
  end if;
  NULL
  
  end proc:
  
###########################################################################

# Examples from the help pages: that works ...

  p := proc( x :: float ) :: float; 2.3 * x end:
  cp := Compiler:-Compile( p, keeptemps=true ):
  cp( 1.1 );

I use Watcom as being shipped with Maple 12 on Win XP, but once had
problems on Win ME (command lines are too short) - and remembered
some stuff (but not all details, they are a bit hairy, and at that
time Thomas Richard, Scientific Computers, was of great help - they
are the distributors for Germany).

May the following helps you to locate or solve the problem:

There is the flag Compiler:-Compile( p, keeptemps=true ), this should
allow you to see the generated source and object files.

The following should show you the relevant settings:

  kernelopts(opaquemodules=false):
  Compiler:-_DumpConfig();

IIRC you can set MAPLE_TEMPDIR to define a specific working directory.

Then more or less showstat(Compiler:-Compile) could be 'understood'
and modified to get a 'personal version' (what I did).

Viel Glück ...

from experience there lot other variants of answers, which leave everything vague :-)

for short: I am not a friend of such things and never do it ...

what also may help is the ratio of views/posts for thread