The problem with fossils is out of all the billions or trillion of living creatures that have lived on earth, we have fossils for only a very tiny fraction. This undersized data set can be misleading, since it cannot represent all the lifeforms. For example, modern humans come in all sizes from petite; gymnastics, to very tall; basketball, to very large in frame and size; heavy weight power lifter. Say science a million years from now found a rare fossils from one tall human who lived a that time. They would be tempted to say that all human, at that time, were tall, since our one data point appears to indicate that.
Next, we find one very short human fossil that is carbon dated to be 10,000 years closer to the present. Science might be tempted to say evolution caused the tall humans to get smaller, when in reality they all existed side-by-side. Fossils are such a tiny percent of all variations of past life, conclusions can become biased, since they will need to be consistent the limited and arbitrary fossil data, relative to the original living data set. The bar keeps changing as new specimen appear.
In the larger sense of fossils defining evolution, consider this analogy. Say we make a design with popcorn, on the ground. It is similar to a mandala. This mural will represent all the life in a connected ecosystem. We come back many years later and most of the design is gone. Some popcorn was eaten by animals and insects, some was washed away by rain or pushed to one side by wind. Others were taken and dropped by birds and much also disappeared via rot. What is left over in terms of the original design is now randomized fossil data, of a once ordered design. The only conclusion science can draw with the fossil data that still remains; randomly positioned fossils, is that there was never any design. Fossils are helpful but also create a conceptual problem since they represent a subset of data that does not necessary has a clear cause and affect connection to the original main living data set.
The third problem is connected to DNA and evolution. It is currently assume that genetic change on the DNA is random and that this change is not directed. The problem with this is random changes on the DNA will lead to more problems than solutions. For example, many people are afraid of genetically modified food even though this is based on a directed change on the DNA. The fear is we do not know the secondary problems than can randomly appear from even this directed change. Any mechanism that adds new bases to genetics in a random way, is even less stable for the present and future. Many sicknesses are due to slight mutations on the DNA. I have yet to see medicine show how a genetic mutation in one of their patients makes them progressive and futuristic. Sickness is more common.
The random assumption appears to work with bacteria, where billions of units are produced in a short time, so if we lose 99%, it is not big deal. We can start from scratch. But random changes in DNA in multi cellar animals, with limited numbers that takes long periods of time to create, would lead to constant extinctions. Life on earth should have stopped with bacteria, if random was the path of the DNA. Directed change on the DNA is more consistent with healthy ecosystems that include larger lifeforms. This is not accepted by science, since science does not know how such a mechanism works. Random is useful when reason breaks down and you need to do something.
Some of this random bias in science has to do with the religion of atheism being anti-religion. Religion assumes an ordering principle for life and evolution called God. The mirror religion needs to do the opposite and assume random, which has conceptual problems.