In this work we predict the maximum amplitude, its time of occurrence, and the total length of Solar Cycle 24 by linear regression to the curvature (second derivative) at the preceding minimum of a smoothed version of the sunspots time series. We characterise the predictive power of the proposed methodology in a causal manner by an incremental incorporation of past solar cycles to the available data base. In regressing maximum cycle intensity to curvature at the leading minimum we obtain a correlation coefficient R \approx 0.91 and for the upcoming Cycle 24 a forecast of 78 (90% confidence interval: 56 - 106). Ascent time also appears to be highly correlated to the second derivative at the starting minimum (R \approx -0.77), predicting maximum solar activity for October 2013 (90% confidence interval: January 2013 to September 2014). Solar Cycle 24 should come to an end by February 2020 (90% confidence interval: January 2019 to July 2021), although in this case correlational evidence is weaker (R \approx -0.56).