We construct a statistical indicator for the detection of short-term asset price bubbles based on the information content of bid and ask market quotes for plain vanilla put and call options. Our construction makes use of the martingale theory of asset price bubbles and the fact that such scenarios where the price for an asset exceeds its fundamental value can in principle be detected by analysis of the asymptotic behavior of the implied volatility surface. For extrapolating this implied volatility, we choose the SABR model, mainly because of its decent fit to real option market quotes for a broad range of maturities and its ease of calibration. As main theoretical result, we show that under lognormal SABR dynamics, we can compute a simple yet powerful closed-form martingale defect indicator by solving an ill-posed inverse calibration problem. In order to cope with the ill-posedness and to quantify the uncertainty which is inherent to such an indicator, we adopt a Bayesian statistical parameter estimation perspective. We probe the resulting posterior densities with a combination of optimization and adaptive Markov chain Monte Carlo methods, thus providing a full-blown uncertainty estimation of all the underlying parameters and the martingale defect indicator. Finally, we provide real-market tests of the proposed option-based indicator with focus on tech stocks due to increasing concerns about a tech bubble 2.0.