The challenge in serious games is to improve the effectiveness of learning by stimulating relevant cognitive processes. In this paper, we investigate the potential of surprise in two experiments with prevocational students in the domain of proportional reasoning. Surprise involves an emotional reaction, but it also serves a cognitive goal as it directs attention to explain why the surprise occurred and can play a key role in learning. In our experiments, surprises were triggered by a surprising event, ie, a nonplaying character who suddenly appeared and changed characteristics of a problem. In Experiment 1-comparing a surprise condition with a control condition-we found no overall differences, but the results suggested that surprise may be beneficial for higher level students. In Experiment 2, we combined Expectancy strength (Strong vs. Weak) with Surprise (Present vs. Absent) using higher level students. We found a marginal overall effect of surprising events on learning indicating that students who experienced surprises learned more than students who were not exposed to these surprises but we found a stronger effect of surprise when we included existing proportional reasoning skill as factor. These results provide some evidence that a narrative technique as surprise can be used in game-based learning for the purpose of learning.Despite the increasing popularity of serious games or game-based learning (GBL), recent metaanalytic reviews have shown that GBL is only moderately more effective and not more motivating than traditional instruction (Wouters, van Nimwegen, van Oostendorp, & van der Spek, 2013;Clark, Tanner-Smith, & Killingsworth, 2015). For example, in their meta-analysis found only a (significant) moderate effect size (.29) for learning in favor of GBL. Likewise, they found a moderate, but statistically nonsignificant, effect for motivation in favor of GBL.GBL influences learning in two ways, directly by changing the cognitive processes and indirectly by affecting the motivation . Preferably both sources should be used to maximize learning. A potential problem with GBL is that the outcomes of players' actions in the