The difference between the two scenarios is not an apples to apples comparison. You are comparing what is probably the least risky investment possible (deferring SS) to an investment that has some bit of risk. The delta in outcomes for deferring SS is practically zero. The outcome delta on a portfolio including equities over 25 years is substantial.
If you wanted to do a more even comparison, model deferring SS while also moving amounts equal to what the early payout would have been from your portfolio bond allocation to portfolio stock allocation. In other words, equalize the growing equity exposure (the risk) between the two scenarios. When you do that, I am confident that deferring SS will beat by a small but consistent amount any projection past your actuarial lifetime and partially close the gap in the case of dying early. The reason is because currently, deferring SS pays a better return than any comparable fixed income investment. SS deferral rates are fixed and they were fixed when we under a higher interest rate environment. Someday interest rates may go up, or SS deferral rates may be adjusted down. Until that happens SS deferral has an advantageous return over comparables.
Most people who retire early with portfolio assets and enter the SS decision years value a de-risking of their withdrawal plans, although almost no one consciously models it that way. Deferring SS is an insurance that a lot of people want. If you don't want the insurance, consider the ways (like the paragraph above) to equalize the risks.
Of course, there's lots of individual situations with health, or cash flow, or spousal/family factors that would trump the above analysis. I don't have a dog in the fight for other people's SS claiming strategies. Just trying to bring a perspective to modelling factors I seldom see addressed when the discussion gets repeated.
I personally am waiting until 70 and I am also aware that the risk reduction of that choice allows for a rising equity glide slope, if I wanted the risk. So far I don't.