That doesn't really matter if engine is only slightly changed with each version, its still so complex even little things can have unpredictable results.
Comparable thing in complexity would be for example chess engine. Top chess players won't go into serious match against computer engine without demanding version is completely frozen , having documentation of every patch, having enough time to practise with patched version before the actual game, stuff like that. Slight change still causes their "sample database" to become redundant, making them unable to prepare and thus at disadvantage.
Similarily , having played a lot of games with old versions of FM does not give anyone ability to figure out whether newest version is fixed.
Sorry, but chess analogy is poor. Both players have a finite number of options for each move. White and black both only have 20 options for their first move for example. Computer chess engine will assign a specfic mathematical value to each move, based on calculating forward through a number of moves, looking at all of the different paths available. (Sure, the number of options soon stack up, but then number crunching is what computers are good at.)
Mechanisms are designed into computer chess programs to indentify favourable positions, whilst calculating ahead in this manner. First thing it will look for is whether it can force check-mate within the number of moves that it can calculate ahead, failing this it will look for opportunities to force the capture of opponents pieces or a favourable exchange, or a strategically superior position. For example, can the computer force you into an exchange that leaves you with weak doubled pawns etc., and when I say force you, I mean that any other option you take is even worse.
And so it goes on. Computer chess is based on clever programming, which is a combination of providing the software with extensive opening theory, knowledge of move combinations likely to lead to either a postional or material gain, end game theory (again, this is well known and it is possible to either force a win, or at least a draw, from a whole host of known end game situations.) Certain laws of the game apply and are fixed. Nothing new has been learnt in terms of end game theory for many decades, if not centuries.
Apart from this, there's not much else, except how long you allow the computer to calculate it's best move.
With computer chess, as with real life chess, there are no probabilities involved. With humans, it's probably the purest form of the meeting of two minds in competition and against a computer the mathematics are all based on calculating finite options and no probabilities whatsoever. Oh and computers don't make blunders, unless they're programmed in.
Of course, I'm talking about computers as per your post, that are programmed to take on the world's best human players.
Much consumer computer software will have some randomness, or intentional stupid moves by the computer, to allow us mere mortals the chance to win, but that's another story.
Football Manager, on the other hand is absolutely full of probabilities, almost everything in the game can be expressed in terms of probabilities, so Football Manager and Chess could not be more different.
As to my experience over the years and whether it translates from one years version of FM to the next, all I can say is that I have played enough of each version and each patch over the years, to conclusively prove to myself at least, that there are modifiers built into the game, outside of all of the known game parameters, including hidden stats that can be seen using FMRTE, that most definitely are used to adjust the probability of you winning or losing some games.
As Liverpool's new manager said, and I paraphrase:-
Some fans will follow their team blindly, whatever happens. (This equates to people who blindly accept what they're told and blissfully carry on playing the game, and there's nothing wrong with that as long as your enjoying it enough to continue.)
Some will question what's happening in their club, new manager's performance etc., but will be open to reason and will carry on with their support of the team. (To me, this fits with people who are prepared to listen to reasoned argument and do there own checking up to see if it's so. As long as finding out that you probably only have maybe 90% control over what happens in your game and that the other 10% comes down to what the game software decides, doesn't spoil your enjoyment of the game or outrage you, you will also continue to enjoy the game.) BTW, the 90% and 10% are arbitrary figures and may not represent the actual values, as I have not attempted to calculate these!
The last group that he describes are the critics who you'll never satisfy, whatever you do, and he doesn't worry about this group, as you can do nothing to change their minds. (To me, those who are determined to bury their heads in the sand, because they don't want to accept that they don't have full control themselves, or for some other reason are not prepared to test things out and find out for themselves, they fit this description from Rodgers.)
I have tried using reasoned argument, based on years of experimentation and observation, to describe what is going on, with many contributors to this and other similar threads, agreeing with my conclusions, sometimes fully and sometimes in part.
For those people who are not receptive to a reasoned debate and cannot provide any evidence to prove the opposite, I say, carry on enjoying this wonderful game with the rest of us, because in many ways it does not matter what you believe, as long as you enjoy the game.
) play à high pressing game, you are more likely to see scenarios like this as when the team dont press as à unit, the approach fails. Here, its not the tactic that fails as such, but the players Will to carry out the orders and the solution would be à tactic change.While this might not be the case Of the OP, its à common reason.