Well, you keep saying that this is "no problem" and the like, but it simply is not so. If the model is already animated and you change the vertex number order in any manner or add/delete vertices then you have a bit of a mess and have to do a bit of re-work to get the model re-animated.

In most cases, adding/deleting geometry (such as adding edge loops, etc) will mess up the UV map (portions or perhaps the entire thing) and force you to rework portions of the UV map.

Reducing the polygon count on a higher polygonal model is not difficult. I do it frequently. However, it is a bit of work in that it is not just the geometry that you have to work with but, as talked about previously, the UV map and, if the model is animated, the animation as well. Yes, there are tools out there to help, but in most cases, even with these tools, you have to do some hand tweaking.

In any case, you are missing the point (I believe). All I am saying is that there are a lot of unknowns about this model. It has a mid-range poly count but looks like a low-poly model. It has a decent sized skin (1024 x 1024), but it looks like a model with a 512 x 512 skin. Because all we can see are the textured shots, we cannot see the wires and therefore we don't know how well the model is constructed.

Yes, I could re-do the topology. Yes, I could rather quickly do a new UV from scratch (it would really only take a few minutes). None of this is the point. The best way to model is to get it right from the start. He showed the model here and I brought up my concerns. That is all. Whatever you can do with a high-poly model, etc, is of no concern to this thread.