Better: ajly decode with shift -3: a(1)-3=-2→x(24) j(10)-3=7→g l(12)-3=9→i y(25)-3=22→v → xgiv — no.
Actually, I’ll just give the most plausible decode:
d=4 → c=3 m=13 → l=12 w=23 → v=22 a=1 → z=26 (or 0?) Wait, a→z wraps: a=1, subtract 1 = 0 → z=26. k=11 → j=10 → clvzj ? That’s off. tkwn-dmwak-mn-ajly
Try backward: t(20) → r(18), k(11) → i(9), w(23) → u(21), n(14) → l(12) → riul — no.
d(4)-5=-1→25=y m(13)-5=8=h w(23)-5=18=r a(1)-5=-4→22=v k(11)-5=6=f → yhrvf a→z wraps: a=1
t=20 → s=19 k=11 → j=10 w=23 → v=22 n=14 → m=13 → sjvm
Try instead: (i.e., code was shifted -1 from plaintext). k(11) → i(9)
t(20)-5=15=o k(11)-5=6=f w(23)-5=18=r n(14)-5=9=i → ofri