Slot machines are a remarkably popular mode of gambling even though they are programmed to make a profit by paying out less money than is put in. One common feature of slot machines, which may increase the likelihood of persistent gambling in the face of this monetary loss, is the near win. This study’s aim was to investigate the conditioned reinforcing properties of near wins using an observing response procedure in the context of a simulated slot machine. In an observing response procedure, participants can use an observing button to produce a stimulus correlated with the availability of reinforcement (S+) or a stimulus correlated with no reinforcement or less reinforcement (S-). The percentage of observing responses made for each stimulus is thought to reflect the reinforcing efficacy of the reinforcer correlated with each stimulus. Experiment 1 successfully tested the procedure with an obvious reinforcer - wins - and found consistently more observing for the S+. In Experiment 2 and 3 the S+ was correlated with near wins, and in Experiment 2 only those with slot-machine experience had consistently more observing for the S+. Experiment 3 increased the probability of wins to enhance the reinforcing efficacy of near wins, but failed to find consistently more observing for the S+, regardless of slot machine or scratchie card experience. These results indicated that near wins are not conditioned reinforcers. However, participants tended to bet more following near wins than losses, which suggested that near wins may instead function as discriminative stimuli.