Terrestrial cosmogenic nuclide (TCN) dating of ice-marginal moraines can provide unique insight into Quaternary glacial history. However, geomorphic processes can profoundly influence the distribution of TCN ages, as boulders are exhumed, toppled and eroded as moraines degrade through time. In turn, quantitative methods which minimise and account for these processes have the potential to significantly improve the chronological utility of the moraine record. Within this context, we focus on a fundamental component of TCN sample selection; are moraine crests better targets than moraine slopes? This question arises due to an apparent dichotomy between model predictions and sampling procedures, as slope diffusion models predict the greatest ground lowering at moraine crests and relative stability on moraine slopes. However, crests are preferentially sampled for TCN, while moraine slopes are generally avoided due to increased risks of boulder instability. In this study, we assess whether spatial criteria can be used to identify exhumed boulders through intensive Schmidt hammer sampling (SH; 505 boulders - 15,150 R) and 10Be exposure dating of four moraines of varying age and geomorphology in the northern and southern Pyrenees. These data show that large, matrix-rich and steep-sided moraines are particularly susceptible to post-depositional erosion, but sampling on the crest is not effective at isolating exhumed boulders. For these moraines, the distribution of exhumed boulders is effectively random, while the probability of selecting an exhumed boulder is comparable for moraine crests, ice-proximal and -distal slopes. In turn, sedimentology exerts a key control on moraine stability, as matrix-poor, boulder-rich moraines stabilise rapidly after deglaciation. In light of these data, we propose that landform stability should be prioritised over spatial criteria and that, in the absence of detailed geomorphic analysis, boulder location should carry less weight in sample selection. However, non-ideal landforms are often priority targets for TCN dating as these may record the retreat stages of the largest glaciers or represent “missing” glacial stages in existing chronologies. For these landforms, we recommend preliminary SH sampling to identify exhumed boulders and to select boulders for TCN dating.