Let's look at what really happens in nature:
We'll start with a salmon stream where salmon have just spawned (and died) and have laid and fertilized 1 million eggs. So we start with one million salmon eggs on the bed of the stream. Not all these eggs will hatch. Many will be eaten; many will be crushed by tumbling rocks or washed up on a beach; and many will be non-viable. Probably less than 10%, 100 thousand, will actually hatch and start to grow and make their way down stream to the sea. Most of these won't make it. They will be eaten, injured in swirling stream waters, or will starve if they can't find enough food. At most, 10 thousand will make it to the sea.
There, they will spend three to five years growing to maturity, at which time they will seek out the stream they were born in. Most of the 10 thousand won't make it to maturity and back to their stream. Many will be eaten, injured, or starve. At most, 1000 will make it back to the stream mouth and start their arduous journey up to the spawning grounds. Again, most will not make it. At most perhaps only 100 will make it to the spawning grounds, about fifty males and about 50 females. These females will each lay over 20 thousand eggs that will be fertilized by the millions of sperm that the males eject.
We are now at full cycle with one million fertilized eggs on the bottom of the stream waiting to hatch.
What this example shows is that the first rule of life is
Survival of the Luckiest. Only those lucky enough to be in the right place and the right time to find food and not be found as food, to not be excessively dashed about by rapids and waves, and to have avoided the slings and arrows (jaws and claws) of outrageous misfortune make it back to spawn. But luck is not heritable. The offspring of the survivors will not have better luck because their parents happened to have good luck. So there is no cumulative improvement in Luck over the generations.
But their is a second, much weaker effect that will help a few of the fish survive - beneficial mutations that make them a little bit faster or less visible to predators, or more efficient digestion, or produce more viable sperm or eggs. This second effect is so much weaker than Luck that it is not noticeable in individual generations. But, it has the advantage, like compound interest, of accumulating over the generations and natural selection will gradually nudge the species toward greater population fitness. So, the vast majority of critters that die (and are fossilized) will actually be quite typical of their generation and 'monster' mutant fossils will be very rare.
By the way, are you aware that you are a mutant? In fact, every cell in your body (and you have about a 100 trillion of them) is a mutant. No two cells in your body have the exact same genome or the same genome as the egg from which you originated. The error rate for base pair duplication when cells divide is (very roughly) about one error in every 400 million base pairs. This sounds very impressive, but when multiplied by the six billion base pairs in your genome, gives one to two dozen mutations for each cell division.
Most of your cells differ from the original egg in hundreds to thousands of base pairs. And yet you have you no problem surviving and reproducing. Almost all such mutations are neutral and have no effect on the organism. To say that most or almost all mutations are deleterious is simply false. Almost all mutations (in the 99.9% range) are irrelevant to the individual and to the population. This provides the broad landscape for the rare beneficial mutations to accumulate and eventually dominate in the population over many, many generations
™ Version 4.2