I derived the following family of binomial coefficient identities rather indirectly for natural $t \ge 2$ (though it seems to hold more generally). I was hoping someone out there might know if this result exists in the literature, or perhaps come up with a more straightforward proof than mine. I paged through the H. W. Gould volumes of identities but didn't see it, though there were a few that seemed similar.
$$ \binom {n}{k}= (t+1 + t n )\sum_{i=0}^k (-1)^i \, \dfrac{ \displaystyle \binom{t + (t-1) n + k }{i} \binom{n+1}{k-i} } {(i+1) \displaystyle \binom{ t+1 + t n + k }{i+1}} $$
Just to check that I've transcribed it OK, let's work an example, how about $\binom{5}{2}$ with $t=3$? That's $n=5, k=2$.
\begin{align*} %% & %% (t+1 + t n )\sum_{i=0}^k (-1)^i \, %% \dfrac{ %% \displaystyle %% \binom{t + (t-1) n + k }{i} %% \binom{n+1}{k-i} %%} %%{(i+1) \displaystyle %%\binom{ t+1 + t n + k }{i+1}} %%\\ \binom{5}{2} &= {} (3+1+3(5) ) \sum_{i=0}^2 \\ & \quad (-1)^i \, \dfrac{ \displaystyle \binom{3 + (3-1) (5) + 2 }{i} \binom{5+1}{2-i} } {(i+1) \displaystyle \binom{ 3+1 + 3 (5) + 2 }{i+1} } \\ %%% &= %%% 19 \sum_{i=0}^2 (-1)^i \, %%% \dfrac{ %%% \displaystyle %%% \binom{15}{i} %%% \binom{6}{2-i} %%% } %%% {(i+1) \displaystyle %%% \binom{ 21 }{i+1} } \\ &= 19 \left( \dfrac{ \displaystyle \binom{15}{0} \binom{6}{2} } {\displaystyle 1\binom{ 21 }{1} } - \dfrac{ \displaystyle \binom{15}{1} \binom{6}{1} } {2 \displaystyle \binom{ 21 }{2} } + \dfrac{ \displaystyle \binom{15}{2} \binom{6}{0} } {3 \displaystyle \binom{ 21 }{3} } \right) \\ &= 19 \left( \dfrac{ 1 \cdot 15 }{21 } - \dfrac{ 15 \cdot 6} {2 \cdot 210 } + \dfrac{ 105 \cdot 1 }{3 \cdot 1330 } \right) \\&= 19 \cdot \left( \dfrac{5}{7} - \dfrac{3}{14} + \dfrac{1}{38} \right) = 19 \cdot \dfrac{10}{19} \quad\checkmark \end{align*}
Thanks.