Interview with the programmer of Micro Bridge


Tomio Uschida answers some questions:

1. What do you think was the reason for winning this year:

Simply a probability. From my experience, Wbridge5 has 50% chance to win, other 5 programs (Baron, Micro Bridge, Q-plus, Shark, Synrey) have 10% chance to win. This year is my turn. Micro Bridge played perfectly against Synrey at final match, but same thing will not happen again.

Have you improved the program?

Yes, but these few years, improvement is very little. Other strong programs seems to have same situation.

2. Have you implemented AI functions in your program or plan to do so?

If you say about modern AI, the answer is no and no plan also. When I started to develop Micro Bridge 35 years ago, the idea was based on the latest AI technology at that time. But now these are classical technology. My program is a mass of rules given by programmer.

3. What do think about AI and Bridge, will it manage to make the program so good that it can beat the world players?

My forecast is very negative, especially in the near future. Many AI game programmer thought that Go is a difficult game for computer to beat pro human player, but Go programm's skill has been rapidly improved. Same thing will not happen in Bridge. Go has huge amount of branch and depth, but is a simple game for computer. One good idea can improve the skill dramatically. But bridge is not simple. It has many elements and one or two good idea will not solve whole problem.