I recently bought a copy of Bombay, the game from Ystari which obviously has the wrong name. I obviously wasn't hung up on the Ystari thing because I disliked both Caylus and Mykerinos, and haven't bothered to try the others. However I do love Bombay.
It's essentially a pick-up-and-deliver game, with the sex appeal coming from the elephant miniatures which can actually carry two little cubes of "silk". I was just explaining to my boss today how good interface design in games prevents you from doing the wrong thing - nobody will accidentally have 3 bales of silk in Bombay. Players have 3 action points on a turn with which they may move, purchase silk, sell silk, build palaces, or, if there's nothing better to do, just get some money.
The market mechanism is quite clever. There are 17 silk cubes in total - 5, 5, 4 and 3 in 4 colours. Each phase you draw 9 of them to place on the markets, and the most common get the cheapest price. This means that if you know the colours - purple is 4 and yellow is 3 - you can buy uncommon colours at good prices. Obviously if you have 2 yellow still on your elephant when 9 are drawn from the bag, yellow is not going to be the cheapest colour and that's quite an asset you have there.
The action points and movements are not quite so inspiring - it always seems to be just a bit too far to go to achieve quite what you want. Restricting yourself to what you can achieve profitably is part of the art of the game. More interestingly, once a player has built a palace on an intersection, if another player passes through there the owner of the palace gains a rupee. Going out of your way to avoid other people's palaces often isn't feasible, so building palaces can be quite lucrative. If movement was easier, that wouldn't work, so the designer has balanced the movements and actions nicely to support the palace toll.
There are 4 cities on the board - Bombay, Indora, Nagpur and Hyderabad - and each demands 3 different colours of silk. With these sets of colours allocated randomly, and the locations at which you can buy each colour allocated randomly as well, the starting conditions of the game are different each time. The best move for the first move of the game depends on how far it is to the silk vendors, how far it is to the cities that demand those colours, and what colours are available. Furthermore, there may be multiple excellent moves - the first player may buy the only yellow, the second player may buy an orange which is in demand nearby, and the third player may rush to build a palace on the intersection the others need to cross. Sometimes there really is nothing good to do, and that's when you can take a rupee.
There are bonuses at the end of the game, and they're extremely important. The person who has the highest combined total of palaces and clients (loose women collected along your travels) gets a large bonus, with smaller bonuses for the minor placings. Players who've sold at 3 of the 4 cities get 4 rupees, and if you've sold at all 4 you get 8 rupees. These bonuses make the palaces and tolls strategy, and the buying and selling strategy approximately equally viable. What combination of those strategies you choose depends on the tactical considerations along the way.
It's common on BGG to say that the game is weak because it's a viable strategy to never do anything and take a rupee each time - in particular Tom Vasel said this in his video review. I mentioned this to Scrabblette who pointed out that if one player did that in a 2 player game it would be equivalent to the other player playing solitaire. Of course I couldn't resist, and sat in bed playing the game. I played twice and scored something like 26 and 29, whereas the recalcitrant player only scored 18. In the 5 player game I played I scored 33. I think one of the players scored less than 17 (which is what you would score by doing nothing, as each player gets one less turn in a 5 player game), so he would have been better off doing nothing... but chances are he got screwed over a few times. Yes, some players might score better by doing nothing, but they won't win.
Finally, Bombay plays astonishingly quickly. In our first game Scrabblette and I were amazed to discover after a few minutes that we were a quarter of the way through. I think our first game took 30 minutes, the 5 player game maybe stretched as long as 75 minutes, and the solitaire games were about 15 minutes including lots of thinking. It's an impressive achievement to fit this much room for thought and strategy into such a small time.
Overall, I'm very impressed with Bombay. It has great bits, multiple ways to win, player interaction, all the things that people say they want in games except for Daleks and an interminable playing time. It's currently rating 7s on BGG, which astonishes me, as this is a truly great design.