Very early on. Long before the game was actually released. So them having stuff that isn't tabletop accurate now shouldn't be any sort of surprise.
Well, I'd say that it's fine to have an imaginative vision for something and not get what you expected in the end. That disappointment is perfectly valid and I'm not going to denigrate it. So I'll throw out the following for posterity and to work out my own experience watching the project make it to release, what my expectations and experience were and so forth. Intended to be reflective and not aimed outward. Two discussion points:
Point, the first: I pretty much saw this coming from a mile away. Much of that comes from following game dev and reading books, class lessons, columns, and conversations within the game design world. It's something that's become formal/academic for a good while now and attitudes about what good design is and how to get a game to ship *and* be successful is recognizable at this point. A flashier MegaMek, that is something very faithful to the board game but packaged up with good art direction, was never going to happen.
There are expectations buried deep within game designers now about what will 'read' to a wide audience (not dumbing down, but what will naturally engage people and keep them drawn into your game). This involves things like how many mouse-clicks it takes to accomplish any action, how information is distilled, what the core game loop is and should be and how long each iteration of the loop should take, what kind of metaphor or story the game mechanics are trying to sell, etc.
Knowing that kind of motivation and background in designers tells me that they're coming from a completely different place than a decades-long Battletech fan. A Battletech fan has the board game and its rules firmly in mind and all natural decisions and expectations follow from that. A video game designer incorporates the board game into his/her decision-making process, but the envisioned experience is never going to stem from the board game unless you're making a strict adaptation like mobile versions of Agricola or Carcassonne. HBS's game was never trying to be that and, while I'm sure they cracked open MegaMek at one point, never trying to be MegaMek either.
I think that's where the conflict in expectations primarily happened. Basically the board-game fan is from Mars and the game designer (Mike McCain, Kiva Maginn, and on down) is from Venus.
Point, the second: I interpreted assurances that we were going to have a board game experience was not an assurance that this *was* going to be the board game. It was that it would be very much a Battletech board game experience as compared to, say, MechCommander RTS or even a Mechwarrior sim. The context was the video game series that are out there. It might be a fine distinction, but it made a lot of difference for me. Especially as I had in mind what Harebrained had done previously. Shadowrun was probably closer to its ruleset than Battletech is, but that's probably an artifact of the natural flexibility of GM-run P&P RPGs.
One of the biggest distinctions of the Shadowrun trilogy was the incorporation of elements from the very successful XCOM, which showed both designers and the audience what a modern turn-based strategy game could do to generate urgency, excitement, and polish. The main thing was how they communicated cover mechanics (directional shields), but there were other subtle bits and bobs that showed off the game's artistic heritage. That was an amalgamation of the tabletop rules and game-design trends in the industry. I fully read that as the context for what Battletech would be in the end and more easily rolled with the changes.