I think the key advice that should be mentioned here, besides "make sure to optimize your run better", is to do research about existing runs, tricks, and discoveries of the game. Especially for a game like Smash Bros., a popular fighting game still played by many players today, who have built upon over 20 years of history in pushing the game to its human limits. The goal of this run is unclear (besides "beat 1P Mode with Yoshi"), but it's clear that research isn't done. If speed is the goal, real-time runs exist that are faster by more than two minutes. If entertainment is the goal, the gameplay fails to impress because even though ostensibly TAS tools are used, the rate of actions per minute is maybe a quarter of that of the playaround runs that are already on the site. The movie misses every essential trick that human players and TASers alike use like z-cancelling, double jump cancelling, or fastfalling, there are no real combos, and the character even unnecessarily takes damage several times. This looks like the playthrough of an average player, never mind a high level player, or a TAS. Runs like this have far more potential, which the author should be aware of. A TAS should showcase a game being pushed to its limits, and the existing material should show to the author what pushing the game to its limits really looks like.