Some way make the pieces much lower entropy than the pawns... then wait a loooong time haha.
edit away my specific example - oh I see, all same material, hmm.
I guess then the quickest answer is to manipulate the board such that temp and pressure vary from rank to rank.
we can use two laser beams to vaporise the pawns.
Queens Tour, 14 moves
Fastest I could find was 12.5 moves, don't know if it's possible to reach that position faster:
In all seriousness, 12.5 moves. The star of the show? "Rooks" again.
EDIT: Well, Uncia_Uncia and I cross-posted with the same idea. How about that.
As for the analysis on Breyer's 50 mover, give me some time to read through it. That one is definitely not your entry-level problem.
EDIT: OK, read the analysis. The relevant bits are sound - location of captures, origins of pawns, promotions. That is the preliminary work of the problem - identifying the presence of a "retrocage" on the left, which is a group of units which cannot be easily retracted back to their original locations.
Here a word on approaching retros: the notion of "retracting" moves makes things much much simpler to analyse. Instead of thinking "to reach this position, the sequence was 1. a b 2. c d 3. e", think, "from the final position, retract the retromoves -1. e d -2. c b -3. a". Which incidentally is neater notation. More on retromoves here.
So your mainlines are -1...Qb4 -2. Rbb5 Rb6 -3. Ra6 Qa5 -4. Rb4 Bb5 -5. Ra4 Bb4 -6. Ra3 Qa4 -7. Ra5 Ba6 -8. Rb5 and -1...Qb4 -2. Rab5 Qa5 -3. Rb4 Bb5 -4. Ra4 Bb4 -5. Ra3 Ba4 -6. Rb5 Qb6 -7. Ra5 Bb5 in normal retro notation.
I think you realise the difficulty of doing this kind of analysis for 50 move problems. This one in particular is a sort of sliding puzzle where the goal is unclear. So the angle of attack here should be, how did all the pieces enter the box / how can you retract pieces out of the box? What configuration of pieces allows you to unlock the retrocage? Then work towards getting the pieces towards that target configuration from the given position.
Haha, yeah this is a rather tricky one(some of the other ones I looked at that you had posted were leagues easier than this one).
Considering the clue that its a 50 move problem, I'm assuming that you can retract back 50 moves without a single pawn move being allowed(or captures, but in this case all captures are also pawn moves), so rather than just finding a way to get all the pieces out of the retrocage(which would show the validity of the problem), I believe here you have to find the fastest way to get out of the retrocage(or at least to get to a legal pawn move) and show that it couldn't have been done faster?
Or, I think another way to do this would be to show the position is valid, and that at least 50 moves is required before a pawn move(aka the line that shows the position is valid need not be the fastest line, and then as a seperate proof show that regardless of how you retract moves, it will take at least 50 before a pawn move is legal)
Those approaches are, practically speaking, equivalent. To unlock the cage (or in general, to enable a pawn move/capture), there is typically a critical position or series of manoeuvres that must be made. This means there is only a single "way" (possibly with different move orders/waiting moves/extraneous moves) to resolve the position and prove legality. Then it's just a question of recognising the key features of this way, and optimise it as much as possible. The resulting sequence should show that >=50.0 moves are required since the last pawn move/capture, and sometimes castling.
I still haven't made much headway on the 50-mover you gave me, but I designed a retro(or essentially a series of retros) on my own:
So, doubled pawns happen quite a bit in chess, but tripled pawns rarely. Quadrupled pawns even more rarely. What about "septupled pawns"? Now you might say that there just aren't enough ranks for that(and you're right), but if we allow the "pawn" on the 8th rank(if you're white) actually be a piece promoted from a pawn, its theoretically possible, soo....
KiloNewton,
Yes it is(even though usually you want to give a potential game that reaches that position to prove it), but what if you put that same shape on another file? is it legal for all 8 files, or is there soemthing special about the E file?
its not possible in all files. e.g. its absolutely impossible for a or h file.
b and g file are also impossible.
i finished the calculation.
its possible for c to f file
Why though KiloNewton? "yes" or "no" is only half the answer, I need reasoning. Why is it impossible?
do a little homework, bro 
I always maintain 10 games =rarely you reach 50 moves
Kilo, I'm sorry you're just not understanding the problem... I already know the answer, so I don't need to do a little homework. Even if I were to "do a little homework", there is no finite ammount of "homework" that can be done to garuntee you come to the conclusion that the position is illegal since any brute-force method will fail due to the fact that chess has infinite continuations, so without a logical explanation as to why it "isnt legal", there is no way to tell if you're just missing something or not. Just as if on the problem you gave, where all the pawns are missing, if you said "there isn't a way to do it in less that 14 moves", that's not a complete answer since you could just be missing something(the way to do it in 12.5).
Kilo, ......
if you said "there isn't a way to do it in less that 14 moves", that's not a complete answer since you could just be missing something(the way to do it in 12.5).
why would i say that?
i've invited naka to solve under 12.5 
see notes http://www.chess.com/members/view/Hikaru
I guess I should have said "if someone says there isn't a way to do it in less than 14 moves".
My point is, without reasoning, you can't say "this can't happen in chess" with any certainty.
I guess I should have said "if someone says there isn't a way to do it in less than 14 moves".
My point is, without reasoning, you can't say "this can't happen in chess" with any certainty.
no one can say that, as you already seen its done with 12.5 moves.
we can argue if its possible by 12 or less.
increase the lower limit
1. 8 moves to capture the 8 pawns
2. you cannot capture in all moves: add the extra moves
3. add moves for diplaced pieces to come back to their squares.
magnetic paws will contradict to go outside the board.
so you may have to break up the magnets from the pawns. (not from the pieces - it'll help them stand on their places)