Noob question on DCF vs. Comps valuation

Here's another perspective - EBITDA multiples are easy shorthand, but they're really only directly meaningful for more or less stabilized businesses. 

Think about it this way: pretend you owned an outdoor theme park and 95% of your annual revenue occurred between Memorial Day and Labor Day. You typically value the business on a multiple of EBITDA because that's what everyone does. In 2020, you get hit hard by a hurricane in April and you have to shut down for repairs during your entire busy season. As a result, your LTM earnings will fall off a cliff and probably be negative until the next summer rolls around. As of that September, multiplying your LTM EBITDA by the typical multiple would result in an enterprise value close to zero or negative. But is the business really worthless? You expect to make the repairs over the next few months, and you'll open back up for normal operations in summer 2021. Pretend you're netting $10M per year with no growth - over the next 50 years, you now expect to make $490M instead of $500M... pretty much a rounding error. You can argue it's worth less (TVM, cost to recuperate earnings, etc.), but in any case the LTM multiple will assuredly be higher than it was before the temporary disruption. This is effectively what happened to restaurants (and theme parks, for that matter…) during COVID

This is the idea behind a DCF - the business is worth money because of its ability to generate a stream of future cash flows, and if the business is still expected to generate income in the future (in line with previous estimates) then the present value shouldn't actually change that much even though the TEV / LTM EBITDA multiple approaches infinity (as EBITDA approaches zero). This is especially pronounced in a near-zero interest rate environment - as the discount falls, today's earnings aren't worth that much more than tomorrow's. We don't even need to look at extreme examples like EBITDA falling to zero; this is the same logic behind applying multiples to "normalized" EBITDA instead of reported. Similarly, you'd pay an above-market EBITDA multiple for above-market growth, synergies, etc. - the idea being that in the long run, earnings will increase to support the valuation even if it looks high based on current metrics.

Your GP is arguing that the long-term earnings potential of the business hasn't been impaired despite a fluctuation in LTM EBITDA, and as a result the historical EBITDA multiple isn't representative of the current TEV. This is entirely sound, internally consistent reasoning, but it does get abused (every business ever will adjust their EBITDA and tell you they're really worth more, and PE naturally wants to use the asset class's illiquid nature to smooth returns) - the job of the buyer is to evaluate on future earnings potential and take a view on what valuation that supports; TEV/EBITDA is just a quick way to reference that but it isn't actually the determinant of value.