The only reason not to buy Lemina would be if he is seen as an injury prone player by our medical staff. Otherwise it would be ridicoulous to pass on a talent like that for 10 mio. We could sell Sturaro for 15-20 mio. and get Lemina for 10 mio. Make some money and get the far superior player. Sturaro is a good player but nothing special. Lemina has the abilities to become a beast of a defensive minded CM or DM. Anybody who has ever played football or followed football should be able to see the difference in potential between the two, small sample size or not.
I would never trade Lemina for sturaro. Sturaro is limited yes in that he doesn't have a wide range of different skills but he has shown very good maturity and reliability at an early age. With him, we know exactly what we will get in the short term and in the long term he could be a beast ball winner as he grows even more. Its not exactly a good analogy but Gattuso was also very limited (would never even make a good run into the box) but he perfected the one or two skills that he had. Now sturaro might never be as good as Gattuso anytime soon but he has potential to grow and is already reliable in the short term. You can't go wrong with sturaro because there is no risk involved. If he develops into a Gattuso (i know its unlikely he will get that good at defending) then you will have a world class player worthy of a starting spot, if not then at the very worst you have a reliable substitute player for life.
In contrast, with Lemina there is a risk simply because he has not proved himself at all. The sample size is extremely important in evaluating a footballer. There are countless footballers who have shown flashes of brilliance and then turned out to be mediocre. We saw that with Giovinco and De Ceglie who both looked very good at a young age and had brilliant devastating games early on much better than anything Lemina has done yet. When they did get a long run of games we saw that their devastating games were flukes and the supposed wide range of skillset that we thought they had turned out to not be there. With a small sample size you just cannot really know whether they have a wide skills set or not.
The key point is this. The process of identifying that someone has a particular skill is by measuring how often they have displayed it over a large sample of chances. How do you identify the skills possessed by a player when you have barely seen him play if possessing a skill means successfully performing it over a large sample? By definition you can't know what skills the player actually has.
Inter learned the lesson with Santon. We learned the same lesson with Amauri. He seemed like he had dribbling skills, flair and the ability to score headers. After seeing him on the pitch for a long time we realized that they were false flags and that the only skill he really had is the heading ability. Heading was the only skill he successfully performed consistently when a large sample was taken, the rest of the skills turned out to be flukes. Looking only at the initial first small sample, you would think that he had Ibra-like skills. The first few appearances are a very weak indicator of a player's skills set.
To be clear, no one is arguing that Lemina does not have goal scoring skills or penetrative passing skills. What I am arguing is that it is too soon to judge. He might demonstrate that he has a wide range of skills and he might not. We do not know because the very mechanism of measuring whether someone has a skill requires a large sample by definition.
By his first few performances the most you could say is that there are weak indications that he has a wide range of skills. With Lemina there is thus the risk of making a hasty generalization from the small sample. With Sturaro that risk is a lot smaller. Given that we all agree that either player should have a substitute role at most (for now) and that a first team material CM should be bought, therefore if we had to sell one to keep the other then we should keep Sturaro.