Dear Thotsaporn, (and Tipaluck and Doron), I looked at your preprint on double deficiencies and wanted to share two findings about the open problem for Av(123). First, there is a polynomial-time dynamic program for f_123(n,x) = sum_{pi in Av_n(123)} x^{DD(pi)}. Given pi in Av_n(123), complement its values by setting sigma(i)=n+1-pi(i). Then sigma avoids 321. Encode sigma by the two zero-one words A_i = [sigma(i)>i], B_j = [j is an excedance value of sigma]. If alpha_i and beta_i are their prefix sums and r is the number of excedances, the valid codes are exactly those satisfying alpha_n=beta_n=r, beta_i <= alpha_{i-1} (1<=i<=n). Indeed, if e_1<...q and sigma^{-1}(q)>k. These are determined solely by A_k, B_q, alpha_k, and beta_q. Explicitly, sigma(k)>q is equivalent to beta_qk is equivalent to alpha_k