The gap between those two numbers cannot be overstated in practical terms.
The Republican typically receives 82-85% of the conservative vote. At 43% conservatives that means the Republican has already banked between 35.2 and 36.5% of the electorate, before the liberal category or moderate category are explored at all. The mathematical margin for error is immense. That's why for decades I have always wagered on Republicans in states with 40ish% conservatives and a favorable betting price. It feels like stealing. Cruz, for example, only received 34% of the moderate vote in 2018. But it was plenty high enough to defeat Beto due to all those conservatives.
I realize the media never mentions this, and that's why so many people refuse to believe it is that simple. But it is indeed that simple.
Another way to look at it is in terms of resistance. I have seen posts here emphasizing that Texas polling has understated the Democrat recently. That is true, but it is an entirely different ball game when a 9 point loss was overestimated at 12, or a 3 point loss was overestimated at 6, compared to a shift of fate 3 points from one victor to the other victor. That ideological wall is so damn close. The more you push the greater the resistance and none of the major categories cooperate toward another 3 point error.
Biden can win Texas. However it would require a huge national margin in the 9-10 point range, combined with an altered Texas electorate dropping to 40% conservatives or thereabouts due to influx of new voters, and some right leaning anti-Trumpers briefly staying home or not describing themselves as conservative. There are occasional examples of that type of thing, like North Carolina 2008. But outlier mode is not something I care about. As a gambler you have to ignore subjective hype, ignore exceptions, and focus on an easy way to allow numerical normalcy to work in your favor.