This game is about optimizing the moves and informations given in order to win (as fast as possible).
The most important thing to understand is during the early game, the goal is to give informations to others players. More the game advance, and less what you check is important, and more the movement is.
During early game, you want to be obvious in what you did. During late game, you want to be effecient.
During mid game, the goal is to optimise the deplacement by trying to gain one turn by moving a card which will link cards of same colors which were one away or rarely two away.
You usually want to peek (in order of value):
- if before your peek you already got a move planned, a tile where certain colors would make that move bad (=peek close to the tile you plan to move)
- tiles you can move
- tiles not adjacent to any tiles you know
- if you search for a tile to move, peek at tiles far from the action. If you search for a tile to lock, peek at tiles close to the action
- if a single color hint is available, peek at tiles close to the action to lock it there
- if before your second peek, you still got no tile to move, try to peek at a tile you could move which would make that tile close to the action
- if you fully optimise, and if you got a move to do before your first or second peek, you can peek at a bad tile (like a tile you can't move) to show to your teammate you already got a pair. In some cases, if you finally don't move your first pair, you will give extra infos
You usually want to:
- group your pairs to make them be adjacent (for the others players to know you got a pair). You usually got two possibilities to move your pair, try to regroup them close to the action/close to unknown tiles
- put low priority on moving tiles which are one or two aways: you got the upside to have the possibilities to 'fill the gap' with the others cards of that type
- if you fully optimise and got more than one pair, put more value on moving your latter pair: if your teammates paid close attention to what you peeked at, they can deduce things from your weird second peek (since you already got one pair)
If a red hint is available, and a red/purple hint is available, you can assume using the red/purple hint for red is straight up bad. That mean you can use red/purple to said it's purple.
For the same reason, if a red/purple/green hint is available, and a red/purple is available, you can use the red/purple/green hint for 100% green.
Putting a hint block the tile which can't be moved anymore. Which mean you shouldn't lock a single card excentred with color rgb.
You only want to lock at tiles:
- where there is at least another tile nearby
- or the color of the card under the hint is known after your hint
Furthermore, locking a single card is not great. You could either never do it, or do it only if:
- so the color is known
- the card is close to the the action
- maybe if your teammate got a timing to lock it or not
Position of locking:
You usually want to lock your pair or trio in one of the extremity. Usually the closer to the action, except if it's confusing. On this case, locking the outside extremity look best. So usually, a locking of an outside extremity mean you got a trio. If you got a pair, and a third tile of that color which is one away of that pair, you usually want to lock the solo tile if the color of the hint is known. If it isn't, that can be confusing, so you use at your own risk. Since that hint is really weird, that give extra informations to the other player. In the worst case if he add another identical color in the following turn, you will just have to fill the gap later with one of the two tiles of your pair.
If you fully optimise, you want to lock a tile at the best position. But, you can use the position of the hint to convey extra information. If you got access to a green/red hint, and have 2 red and 1 green, you want to first lock your two reds, but give a place to put a card next to it, which would touch both your pair of red, and your green. That way, whatever card your teammate put next to it, it will be good. So, if you don't respect the previous rule, that for me mean you are in this specific case. And so depending of the move and position of lock done by the you, the other player can deduce differents things.
Example of a 2 player start (outdated, but still good):
Here is my personnal start for 2 players if the other player doesn't interfere with what I do and nothing special is going on: https://www.zupimages.net/up/22/07/v4ej.jpg . (it's a great one for effeciency, but far from optimal)
That start is done to first, have an okay start if you first four peeks are differents (path 100> 80V 46< 11^), that will let you have two pairs next to each other (or one triple) turn 3 all the time.
Here is the same document for 2players https://i.ibb.co/tDCSKwV/yokai5-2.jpg done to create two differents '+' for both players. Depending of where the first player center his '+' (c3 usually, sometimes d2), the second player will center his '+' in differents position (usually b1, sometimes c4, rarely b3).