Suppose you are evaluating a new strategy from a quant trading analyst, who claims to have discovered a new high$-$frequency equity arbitrage opportunity uncovered using multi$-$tiered recursive neutral networks $($RNN$).$ Reflecting on what we have learning from ML$/$AI in investments, what should your first question be$?$ A$)$ how many hidden layers does the network have? b$)$ what data went into the model c$)$ what is the in$-$sample performance?