There are several suggestions:

1. Only one simultaneous bid for both players.
2. Play only with letters on board (no more letters inbag). Board will be bigger (7x7, 8x8, 9x9). Game will be pure print-and-play game.
3. Remove doubles and triples fields.
4. Remove will letter
5. Change score counting: score for move =SumLetters * LengthOfWord or SumLetters * (LengthOfWord-1)
6. Change board to hexagonal board (37, 61 or 91 fields)

Some ideas ?

I love the simultaneous bid idea, that would save a lot of time at the beginning of the game and force people to make a sensible bid, rather than just the lowest possible one.

One interesting question is what the exact effect of the bids on the game is.

My opinion would be that the player who chooses the higher number goes first, while the player who chooses the lower number receives the average of the two bids in compensation.

Another possible improvement that has occurred to me would be to connect the 1-2 and 2-1 points on the 2x word score corner, but not on the 3x word score corner.

Personally I think that the highest bidder pay what they bid is right, both players getting one simultaneous bid and highest bid pays that amount ensures both players should bid what they value the grid at. An averaging of bids would increase the likelihood that the contest would be effectively over before a word has been made.

I'd also like to be able to bid a negative number, I did actually once lose a game because I bid 0.5 and my opponent passed and it was clear that the grid had a negative value.

My main suggestion would be reforming the auctions, the other changes could be interesting as variants. When the grid is worth 170 increasing 1 point at a time is slow. Note that it currently correct play to increase by 1 each time to avoid bidding patterns being exploited at the imperfect human level. A perfect player getting first bid should bid 0.5 less than the theoretical value of the grid but I doubt that any perfect GWG playing entity exists currently.

Why not make bids integers, if the player are within 0.5, then a draw would be reasonable this isn't like the game is very drawish.

When you can theoretically play perfectly and lose something is wrong.

Would be nice to have a way of seeing how much your planned word scores and what letters appear in the place of it. Too often games are lost because the highest scoring play allows a big reply. It is possible to do this with pen and paper or large boggle, but it is would be nice if that wasn't necessary to look ahead.

what would happen if both players bid the same?

Same bid and whoever is designated white by the server should start, or other random method.

I am trying simple word game without bidding, scoring. It could be played as print-and-play:

• First move is “SWAP” move
• Letters are in 5 colour groups. Player score one point if she has majority stones in group.
• Game is played as GWG game - when letter is taken, this field is consider as bridge between letters.

Variants: 1. smaller board with 61 letters2. 7 groups instead of five3. Game with numbers instead of letters (image)4. Change rule for pick letters: Choose any letter in one for create word. For example I could play GATE in first row.