Easy OTB randomization of Chess960

Use the poker cards to draws in random the order of pieces.Ace for number one.While the numbered cards are represent the pieces order in regular chess set up.

It can be much simpler.  After random setup of the first rank:

1. If you have same-colored Bs, swap the left side B with an adjacent piece like a1 with b1, c1 with d1 or e1 with f1.

2.  Rearrange K and Rs to R-K-R order if necessary.

This does not make every setup equally likely, because if the right Bishop is on h1, the probability that the left Bishop get to a1 is twice as large as the chances it ends up on g1, because it combines the chances of being on a1 or b1.

I see, let's modify the procedure a bit: Write "LB" (light Bishop) and "DB" (dark Bishop) on bottoms of two pawns.  If the LB lands on a1/c1/e1/g1, swap it with b1/d1/f1/h1.   Same way for DB.  Is this OK?

Under my system, pawns are marked simply with "K,Q,R,R,LB,DB,N,N".  No calculation is needed. 

The probability to get Bishops on a1 & b1 is still twice smaller than that they will end up on a1 & d1. The latter will occur when (LB,DB) initially is (d1,a1), (d1,b1), (c1,a1) or (c1,b1). The former only when it is (a1,b1) or (b1,a1).

1. Put aside two B pawns and devide the remaining six into two groups of three pawns.

2. Add one B to each group; now we have two groups of four, each including one B.

3. Shuffle and distribute one group on a2/c2/e2/g2 and the other group on b2/d2/f2/h2.

Yes, that would work. It is debatable whether it is simpler/faster, though. You would first have to look under the majority of the Pawns to find the two Bishops, then a first round of shuffling to randomize the remaining 6 (most of which you have now seen), then again two rounds of shuffling of the groups of 4 to obscure which were the Bishops that you added...

Compare that to a single blind shuffling of all 8, and only in 43% of the cases an additional Bishop swap. You find the Bishop that should be swaped as a side effect of the final looking under the Pawns to see what goes where, something you have to to in both methods. It does not require any separate action. Even if you fail to remember where it should go, you can just look under known Pawns (those before the left-most Knight and Rook) again. You don't have to search for those by looking under all Pawns. (Imagine how both methods would compare on a 32x32 board!)

Your method is conceptually simple, though. And, more importantly, it requires the same sort of hidden inscriptions on the Pawns. So there is every incentive to prepare the Pawns with K, Q, R0, R1, N1, N3, B and blank inscriptions, and then OTB players could use them in the way they like best.

yep hgmuller have introduced a good and speedy enough way to randomize the pieces without having to reshuffle the pawns for an invalid result etcetera. the benefit from the dices I first suggested are two. No external equipment, which you can lose or fail to find when needed. also about ten to fifteen dollars cheeper. the benefit of the dices are also two. you save 30 seconds approximately to set up. also you need not worry about opponent notice differences on the pawns.

I mean the benefits from the pawn markings in my first part of the rant.

Easiest method I can come up with that gives equal chances to each 960 position:

get the 4 knights and write 1,2,3,4 on the bottoms. Put em in a bag, pull one (j), and stick one of white's bishops on the jth dark square from the left. Put knight j back in the bag, pull again (k), and stick white's other bishop on the kth light square from the left.

Then get 6 of white's pawns and write K/R on three of them, N on two, and Q on the last. Put them in a bag, then pull them one by one, arranging their corresponding pieces from left to right on white's back rank. Just make sure the king is between the rooks, of course, and you're ready to play!

Write on the bottom of the white pawns: Q B N N

Write on the bottom of one of the black pawns: B

Shuffle and place 3 blank and the B pawn for each color on the respective light and dark squares, determine the placement of the two bishops.

Shuffle the Q N N and 3 blank white pawns and place them on the remaining squares.  Place the Q and Knights, and put the king between the rooks

Is there anything wrong with this very simple method (requiring just one dice)? It comes from p.322 of the excellent 'A World of Chess' by Jean-Louis Cazaux & Rick Knowlton (2017). The book covers the development and variations of chess "through centuries and civilisations".

Certainly nothing wrong with it, other than it requires you keep a die with your chess set.  I was responding to the prior conversation about a method that doesn't require additional materials (you could prep your pawns ahead of time).

I'm a lot more likely to have a phone with me than dice.

The weird assumption around here is that playing OTB implies an OTB randomization method. There is nothing wrong with a mobile app providing the piece setup which you can implement on your OTB chessboard. Happy hunting (for the app)!

Btw, I predict that OTB chess will survive but that future OTB chess sets will all come with electronic tools for things like "chess960 randomization", "game recording" and for "post-game analysis".