This is entirely meaningless, of course, unless these traits actually result in reproductive success, regardless of what they conduce to, and regardless of the environmental context in which they attempt to reproduce.
Hmmm. Let's see.
No. It's quite meaningful. Traits that confer success have a statical propensity increase in proportion over time relative to the advantage they provide. It's a central insight that is just as much important for evolutionary algorithms used in engineering and computer science as it is applied in biology.
Here's natural selection in a nutshell:
1) There exists a population of replicating things with individuals varying in the traits they possess.
2) The traits in question are both heritable and confer different probabilities of reproductive success*
3) In each successive generation, traits that confer greater probabilities of reproductive success are likely to be found in greater proportion.
So to take a simple example of natural selection in action, when antibiotics are introduced into a population of bacteria those that have traits that provide resistance (given no overwhelming drawback) are more "fit" since they have a trait that will make it likely have reproductive success relative to their peers. All things being equal, their peers stand a certain % chance of antibiotic related death they lack. Hence, over successive generations we'd predict that this trait is likely to become more frequent in the population. Overtime, natural selection will shape this population to have antibiotic resistant traits it initially either lacked or only had in small proportion. This quite clearly isn't tautological. Since this is an easy case it's easy to look at the antibiotic resistance trait as defined here and conclude it was more fit regardless of what circumstances ultimately befall the bacteria who possessed it. We know the trait is more fit because we know it can negotiate an environment full of antibiotics better than those without can. In short, the antibiotic resistant bacteria is better adapted for an antibiotic laden environment, ergo selection will favor it.
In retrospect it seems real obvious, but this mechanism was only occasionally nipped at prior to Darwin and Wallace. They were the first to put on a convincing case that this mechanism has shaped the biodiversity we see to a significant extent by laying out the case for adaptation existing in nature.
Of course, what actually is fit in a given environmental context isn't always easy to know beforehand. The question of how we know a trait is
adapted - how we distinguish between fitness and drift accounting for changes in the frequency of traits to the extent we can - is a relatively complicated one that goes beyond the scope of what needs to be established here. What's important for the claim made here is that it isn't assumed whatever survives is that which is the most fit anymore than we should assume whichever baseball team wins the game is the team best built to win games. This answers the tautology charge.
*Reproductive success can get tricky to define, but it means something close to having the greatest amount of viable offspring that will also competitively produce viable offspring. 