With this mod, only 2 out of the 3 major ship upgrades can be purchased. Certain people don't like the idea, when reading the mod's description. But many changed their mind after playing the mod, and getting a better understanding of why I did it this way. In this article, I try to give some answers and explanations to the reasons I've been told why people don't like this particular design choice (when they haven't played with the mod).

But... it arbitrarily increases the risk during the approach?

No, it absolutely doesn't. What the ship upgrades do, and how they are installed, are two aspects that are deeply interconnected, and thus modified in concert. With the current system, there is a 50% of failure when an upgrade is not installed, and 10% for the upgrades that are installed. If all 3 upgrades could be purchased, I would definitely have increased the probability of failure. Notably, the chance of failure would have been 30% when an upgrade is installed. So it would be just as risky (or a bit more), but spread evenly over 3 events, instead of being more focused on a single one.

I believe this is important to make the approach more tense, by combining events with low and high chances of failure, instead of a mid-range flat probability for each event. If the shield is not upgraded, the player will go into that section knowing that there is a very real risk, especially when Joker says stuff like "No time to upgrade now, we're going in!", and can be actually relieved if nobody dies. And the next section with upgraded weapon will give a much larger sense of security, but may still create a bad surprise.

But.... the approach is still too random and risky?

However the ship upgrade system is set up, the approach is indeed the most purely random part of the mission, since Readiness scores do not matter. The chance to lose at least one squadmate is about 60% over the 3 different events, which may seem relatively high on paper. Losing more than 1 squadmate is possible but unlikely (this is also an advantage of having only 2/3 upgrades installable: I can tweak the chance of losing 1 person without increasing too much the odds of losing more than one). Someone told me they would be fine with 40% chance of losing someone during the approach, but won't even try the mod with 60%. Let's be aware that such a difference would have absolutely not impact whatsoever, in 4 out of 5 playthroughs.

Even if the approach is more random, I believe that having a significant chance of losing someone there is important, as it provides a bit of a wild card on the team on arrival, and may push the player to react and adapt their strategy (people reported in comments that they ended up in this situation, and found it a very positive thing).

That said, since the approach events are very random, I felt it would be good to give the player at least some possibility to reduce the odds against specific squadmates they might particularly care about (which would not be possible with flat probabilities across all events). 

But... it breaks immersion because there is no lore reason for it?

The ship can only support two major upgrades because of their weight/power requirements. I honestly do not see how this is can be problematic to accept, unless vanilla is taken as an absolute reference of what the ship can or cannot support. But this is a mod, it modifies things. And this is not modified in way that is particularly unrealistic.

Already in vanilla, the whole ship upgrade system is a gameplay element that requires a whole lot of suspension of disbelief. If someone can accept that they can send a few probes (or a single one, with No Minigames) to a planet and completely deplete it of 4 specific minerals, somehow teleport back and store all these megatons on the Normandy, then click a button to instantly transform those into a brand new plating for the ship, right in the middle of empty space. Then they surely can accept that maybe the Normandy can only support a limited number of heavy upgrades.

But... it prevents Shepard from doing everything they can to fully prepare for the mission?

Again, this only makes sense if vanilla is taken as the absolute reference of what it means to be fully prepared. Why is "being fully prepared" equal to "installing 3 major upgrades"? Couldn't Shepard have found other ways to improve the ship? Install additional upgrades? Probably, but it's not in the game. Does the fact that Shepard cannot install more than 3 major upgrades also mean that Shepard is artificially prevented from fully preparing for the mission, already in vanilla?

With my mod, Shepard doing everything they can, and fully preparing the ship for the suicide mission, is equal to "installing 2 major upgrades". Not because Shepard is suddenly incapable of clicking three buttons, but because (in the ME universe as modified by my mod) the ship is physically incapable of supporting more than two. Shepard still has complete agency over their own actions, but no magical ability to change the environment at will. And with my mod, the properties of the environment (in this case, the Normandy) are simply different than in vanilla.

By the way, my mod also changes what it means for the squadmates to be fully prepared. In vanilla, "fully prepared" = loyal. With my mod it is more complicated. 

But... it still removes a choice from the player?

I would argue that it removes an illusion of choice, and gives an actual choice instead. In vanilla, the player is given the "choice" to install these 3 upgrades, or to not install them. But this is not really a choice since there is one option that is obviously better than the other. Does anyone ever stand in front of the upgrade panel, pondering on that choice, and wondering whether they should upgrade or not? No. The only reason someone does not install these upgrades in vanilla is either 1) they didn't pay attention, or forgot about it, 2) they didn't have enough minerals and couldn't be bothered to farm for more, 3) they intentionally wanted to kill a specific squadmate. In terms of game mechanics, none of those are interesting reasons for failure.

With my mod, the player is forced to choose which upgrade to give up on. I'm not saying this is the most interesting choice ever. But it is a real choice, with real consequences, and without a single option that is obviously better than all others. If the shield is not upgraded, and Tali dies because of it, Shepard (and the player) will have more responsibility than if it was all purely random. A kind of Virmire-lite choice where, if you go help Ashley, Kaidan would still have a decent chance of surviving, and Ashley would still have a small chance of perishing.

But... Shepard has no way to know which squadmates are more at risk for each event, this encourages metagaming?

This is a fair criticism, and if I had designed the whole thing from scratch I would have done it differently and tried to avoid this. But I have to work with what's already there, and with a limited supply of my time.

However:
1) The whole vanilla Suicide Mission is very, very metagamey. Once the mechanics are known or even partially spoiled, the player makes decisions based on the known outcomes rather than the situation. So yes, my mod has one aspect that may be seen as somewhat metagamey, but overall it is way less than vanilla. And even then, the player still doesn't decide directly the outcome, because it is not deterministic.
2) Since the approach has a strong random element, I wanted to give the players some possibility to give more protection to specific squadmates that they particularly care about... even at the price of a bit of metagaming.
3) This can be completely ignored by the player, and then who is put more at risk during the approach will just be essentially random/unknown. It is unfair to blame the mod over the fact that I disclosed the precise mechanics for whoever wants to read it in detail.