On this page, we will briefly look at how to solve a ko problem in MadLab. Open up a new 9x9 board () and enter () the following position (Source: James Davies: Tesuji, p. 45):

The question is whether the five white stones marked 'a' can be captured? Try right-clicking 'a', and choose "Capture". MadLab starts thinking (and soon reaches lambda-order n=5). This does not seem to work: so stop the search again (). We will now try to donate Black some external ko-threats before searching. Activate the menu "Options" --> "Ko options...", and a window pops up asking about external ko threats. Try setting 2 ko-threats for Black. Now solve the problem again (by hitting the   button). This time the goal is acheived quickly (for lambda-order n=4), and MadLab plays Black '1' in the figure below:

Try answering black '1' with white '2'. Hit the  button again, and MadLab moves at black '3'. You now counter-attack at white '4'. Hit  again, and now MadLab will respond at black "5". It may seem that White wins, but if you capture the black stone '1' (by playing at the intersection below '1') and hit  again, the program recaptures and informs you that "Black uses 1 out of 2 remaining external ko-threats". By means of the ko-threats you have donated, MadLab is able to solve the problem. MadLab keeps track of the remaining ko-threats at all times, and donating 2 ko-threats in this problem works, but 1 is not enough. To see whether a problem has ko-flavour at all (before determining the minimal number of ko-threats needed), you may start by donating plenty of ko-threats: choose "infinite". Unlike the life & death solver GoTools, MadLab cannot automatically determine the minimal number of ko-threats needed: you have to do this manually by trial and error (but besides this, ko is implemented in pretty much the same way as in GoTools). It should be noted that MadLab uses the super ko rule.