I remain skeptical anyone really knows what openings players should play at what stage of their development.
I suspect also that the real answer, if there is one, will depend on a person's talents and preferences as much as any general principles.
If the Sicilian isn't working for a player, he can always try a different one, as Reb did. Though I wonder how many official chess coaches and authorities would have recommended Alekhine's as the proper remedy for a class player getting poor results with the Sicilian.
But the Alekhines does well in the Slow Chess League -- it's rarely played but performs better than all other KP defenses:
Games %win %draw %loss
55 [0.42 0.07 0.51] Alekhine's
My guess is that greater opening familiarity bodes well for class players too. Reb's success with Alekhine's might have been because his opponents had rarely played against it. (I've never played against the Alekhine's in a rated game, and not much more often in speed chess.)
sicilian defense as a main weapon

I think so too Ipcress. The introduction to Alekhine Alert! is basically just IM Taylor saying that Alekhine's Defense is good because most people won't play the Modern Variation against it.
He says in the intro that in his games he was seeing things like 2.Nc3 and 2.d3 and only got the Modern variation 18% of the time. Some direct quotes "Unless you face Anand, your opponents will probably throw everything at you except the Modern Line" "These statistics do give a very good reason to play the Alekhine's Defense: most of the time you will face not very good lines that you can equalize cleanly against, or even get an advantage right out of the opening!"

I played the alekhines defense for several years while working my way from C class to A class . There was much less theory to know than playing the najdorf as most class players in the South in the 70s were playing the 4 pawns attack and those who didnt usually chose the exchange variation . I scored much better with it than the sicilian at that stage of my chess development and it didnt require nearly as much work theory wise .

GM Lev Alburt tells the story of his Sicilian days when he lost a critical game to Karpov in Alburt's pet line. He decided to chuck the Sicilian and specialize in the Alekhine's. He did well with it.
Fischer was willing to employ the Alekhine's twice against Spassky for the World Championship, scoring 1 1/2 points.

What I would make of your statisical aggregation is the following. You don't have enough data points to draw reliable conclusions (not your fault). Not even close. It's interesting, and thank you for pulling it together. I would like to see the same thing done with a million or more games. Of course, there are always other factors to skew the results, and one cannot assume cause-and-effect.

Chicken: I agree.
But it's a handy dollop of data for class play and it is better than the purely anecdotal hand-waving which otherwise is the common currency in these discussions.
There are big databases out there but those are mostly oriented to 2000+ players. I'm interested in what works at the class level.

Several times I've been curious about what openings the greats played when they were mere class players. Did they stick with the 1...e5 program? It seems hard information to come by.
Today I looked up Kramnik's first published game. It was the Black side of one sizzling Sicilian, a Fischer-Sozin Attack, with kings castled on opposite sides. White's attack looked unsound but fierce. Kramnik managed to balance kingside defense with a queenside counter-attack and won, though it looked hairy in spots.
Kramnik played this when he was eight or nine. It's at least 1800 level play, probably higher.
I think NM Reb's general, maybe controversial, statement was something like "lower rated (with some rating no. mentioned) players might not want to make the Sicilian their major defense against 1.e4."
There probably are some compelling reasons why this is true. A main one might be that lower rated players may not have the experience to understand the subtle strategic plans, positions and timing of moves of the very dynamic Sicilian. Especially when compared to the Ruy or even Alekhine which seem strategically more straight forward.
Making it even tougher, lower rated players who can often beat up on other lower rated players with the Sicilian eventually meet up with players who really know the opening and get beat consistently without being able to figure out why. Not infrequently they figure, because the Sicilian is an "attacking" defense, they didn't push hard enough which probably only makes things worse.
Sure anybody can play the Sicilian and maybe win with it. But unless one is willing to put in the significant time needed to understand and/or has competent coaching to explain all the strategic intricacies, it may be more helpful to a lower rated players progress to play the Sicilian less often and spend most of their time on something strategically "easier" to master as their number one vs 1.e4.
Interestingly, I saw a press conference from the recent Tata, where Seirawan (with Giri I think) mentioned that he thought today's GM best reply to 1.e4 might be ...e5.

Several times I've been curious about what openings the greats played when they were mere class players. Did they stick with the 1...e5 program? It seems hard information to come by.
The Play Magnus app supposedly plays as he did at various ages. I don't know how accurate that is, but I thought I hear it actually imitates his former styles. I don't know more.
Interestingly, I saw a press conference from the recent Tata, where Seirawan (with Giri I think) mentioned that he thought today's GM best reply to 1.e4 might be ...e5.
"Best" in what sense? Most likely to win? Draw? Does it take into account a cost-benefit analysis of having to stay abreath of the Najdorf theory changes each month?
Interestingly, I saw a press conference from the recent Tata, where Seirawan (with Giri I think) mentioned that he thought today's GM best reply to 1.e4 might be ...e5.
"Best" in what sense? Most likely to win? Draw? Does it take into account a cost-benefit analysis of having to stay abreath of the Najdorf theory changes each month?
It's doubtful most GM's think in terms of a Black defense most likely to win. GM vs GM now days anyway since a defense most likely to win probably also has an uncomfortably high potential for loss. I'm guessing Seirawan's thought is probably easiest path to equality balanced with opportunity for an advantage.
He mentioned it as a side note to the Yifan - Giri post game I believe as Giri played ...e5 rather than his usual ...c5 vs Yifan. Apparently Seirawan had discussions with some other analyst about e4 defenses at the event, though I haven't seen it.

It's the best defence to e4 IMO. It unbalances the game and has so many sound variations it's difficult for white to prepare against. In other openings I have tried out, White can often deny black an easy clear plan and black has to settle for a draw. I appreciate that it is so mainline and if you have no time to study the game then Alekhines or Scandinavian might suit but long term they are inferior and in today's world were you can often analyse openings and opponents so well it is a short term view.
Let's say you are correct. A lot people agreee with you.
Then the next question that pops into my mind, is what subset is best for Black when employing 1...e5 against 1.e4. Obviously, White has a partial say as to where the lines go, but are you making an assumption that the play begins 1.e4 e5 2.Nf3 Nc6? Or are you making another assumption? Just curious. What if you KNOW that White is willing to play the Ruy? Best path for Black?

Interestingly, I saw a press conference from the recent Tata, where Seirawan (with Giri I think) mentioned that he thought today's GM best reply to 1.e4 might be ...e5.
"Best" in what sense? Most likely to win? Draw? Does it take into account a cost-benefit analysis of having to stay abreath of the Najdorf theory changes each month?
It's doubtful most GM's think in terms of a Black defense most likely to win. GM vs GM now days anyway since a defense most likely to win probably also has an uncomfortably high potential for loss. I'm guessing Seirawan's thought is probably easiest path to equality balanced with opportunity for an advantage.
He mentioned it as a side note to the Yifan - Giri post game I believe as Giri played ...e5 rather than his usual ...c5 vs Yifan. Apparently Seirawan had discussions with some other analyst about e4 defenses at the event, though I haven't seen it.
A factor would be whether it was must-win (tournament at risk), or whether a draw would be sufficient (ala Carlsen v. Anand recenty).

When I returned to chess, I was shocked to discover how effective the Berlin Wall and the Petroff's had become in obtaining draws for Black, thus reducing the power of the almighty Ruy Lopez for White.
I sensed a disturbance in the Force....
Of course, that's at the GM level and I doubt it means all that much for class players.
Looking at the Slow Chess League archive of 5500+ games I see that KP games outnumber QP games about 2:1. The KP games break down:
Games %win %draw %loss
1192 [0.51 0.10 0.39] 1. e4 e5 King Pawn
996 [0.46 0.10 0.44] 1. e4 c5 Sicilian
425 [0.48 0.09 0.43] 1. e4 e6 French
The Slow Chess pool seems like a reasonable cross-section of chess.com players with standard ratings of 700-2300 with an average rating of 1428.
Looks like the Sicilian players score significantly better than the 1...e5 players and slightly better than French players.
Make of that what you will.