Fisher Yates Shuffle for cards or any number of items

[bplus]

As discussed in Ranking Poker Hands:

Code: (Select All)

TopN = 52
ReDim n(1 To TopN) 'repeatable for ref
For i = 1 To TopN
    n(i) = i
Next
For i = TopN To 2 Step -1 ' Fisher Yates Shuffle of N Items
    Swap n(i), n(Int(Rnd * (i) + 1))
Next
For i = 1 To TopN
    Print "  "; i; "-"; n(i); Chr$(9);
Next
Print

At maximum you need only swap n-1 items!

[PhilOfPerth]

As discussed in Ranking Poker Hands:

Code: (Select All)

TopN = 52
ReDim n(1 To TopN) 'repeatable for ref
For i = 1 To TopN
    n(i) = i
Next
For i = TopN To 2 Step -1 ' Fisher Yates Shuffle of N Items
    Swap n(i), n(Int(Rnd * (i) + 1))
Next
For i = 1 To TopN
    Print "  "; i; "-"; n(i); Chr$(9);
Next
Print

At maximum you need only swap n-1 items!

Is this faster than 

For a = 1 To numtiles
    swop = Int(Rnd * numtiles) + 1
    Swap tiles(a), tiles(swop)
Next
To me, it looks the same, only in reverse order    

[bplus]

Nearly as fast but you are doing one extra swap.

But more importantly, it is less efficient, you are swapping each time with a random from whole set which can cause ripples or waves that upset a normal distribution. The wiki will confirm this but the math is quite wonky.
https://en.wikipedia.org/wiki/Fisher–Yates_shuffle see Naive method which is yours basically.

[PhilOfPerth]

Thanks bplus. 
I thought there would be some subtle difference, but couldn't see what it was.
But I think I'll stick to the "Naive" method - it's simpler, and, for me, simplicity is of the essence.

[bplus]

Thanks bplus. 
I thought there would be some subtle difference, but couldn't see what it was.
But I think I'll stick to the "Naive" method - it's simpler, and, for me, simplicity is of the essence.

It's subtle and I can relate to sticking with things I understand.

Certainly we've all bigger fish to fry!

Thanks for checking it out and you are not alone, Steve wasn't buying the math stuff either when I posted at the other forum. Oh well... carry on