The point is that the data, as Merric observed, is misleading. The base "attack progression" was of vastly different importance to the overall attack power of a character between different editions. While the data itself can be construed an attempt to make an objective "neutral" comparison between editions, the stated purpose and conclusions drawn by the author are very misleading. It does not serve really any useful purpose, certainly not the purpose the author is implying. It's like comparing the average points scored in a professional basketball game to the average goals scored in a professional Hokey game. More points are scored in a basketball game, therefore it must be...? better? Maybe for some people. More exciting? Again, maybe for some. Require more attentive score keepers? Probably? Have better Athletes? But there are more factors that go into these conclusions than are represented in my stats.
Moreover, I would argue that the author is putting value judgments on editions based on faulty logic and misleading data with quotes such as: "the Fighter's combat efficacy is relatively unchanged from AD&D through 3.x, but is adjusted significantly downward in 5th edition". As has been pointed out, the contribution of BAB to "combat efficacy" varies quite a bit between editions, thereby rendering the first part of the quote about the similarity between AD&D and 3E very questionable at best. Furthermore, BAB is measured against AC most of the the time, and I see no data about how ACs have varied or stayed the same from edition to edition. I'm reminded of the old saying: "There are lies, damn lies, and statistics". Anyone can put together a spreadsheet with numbers that are technically correct and seem to support a desired conclusion, but the real, useful conclusion is usually a bit more nuanced and harder to see at first.