We employ principal component analysis to identify components of subject's strategic IQ in the following three classes of games: The two-stage dominance-solvable game, Chen, Huang and Wang (2013)'s simultaneous spatial beauty contest game, and the first-mover spatial beauty contest game. Parallel analysis retains the first five principal components (PCs), which account for 56% of the total variance of subject's normalized expected payoffs for each of the 33 games. We interpret these PCs as five strategic IQs: The first SIQ indicates subjects' abilities to perform backward induction and it is also the common g-factor that can predict subjects' performances in most games. The second SIQ could be interpreted as subjects' abilities to perform high dimensional backward induction. The third SIQ controls for subjects' attitudes toward risk. The fourth SIQ reflects subjects' beliefs about social preferences. The fifth SIQ measures subjects' accuracy of higher order beliefs about others.