So what IS the one thing I would change? You may have guessed it already, it's the two-upgrade limit. However, unlike other people who have posted their objections to that limit here, I do have a slightly different take on it. I can certainly tell you, I absolutely hated the idea of the two-upgrade limit when I first read about it on the description page, and almost decided to not try the entire mod at all based on that feature alone. The fact that there is still a significant possibility of survival even if the upgrade is not installed is the ONLY thing that made me feel okay to get over that detail and install it after all. (Spoiler alert: I'm really glad that I did.)
But once I started charting the numbers on a spreadsheet and figuring out how they interact with one another, I started to understand why you might have done it this way, instead of just implementing an elevated probability of death-despite-installation across the board... I'm not confident enough in my grasp of probability theory to say that I have the right idea of how your mod actually crunches the numbers (not to mention the fact that a word-based description simply cannot get precise enough to truly communicate everything going on at a mathematical level). However, trying out multiple possible ways of calculating the numbers as you described them, it suggested to me that creating unevenness, not just a simple random element, in the survival / death checks is what creates half the tension and suspense of your version. And it was that suspense that comes from genuinely not knowing what is going to happen next that I guessed (since this was at the point I hadn't played the mission yet) that was going to create the correct emotional atmosphere around the approach to the Collector Base... a feeling of suspense that is frankly absent from the vanilla SM, if you've been paying even a little bit of attention through even a first playthrough.
My guess turned out to be right. It turned out I was on the unlucky side during the approach, as not only did my not-installed upgrade result in a death, but one of my installed upgrades also ended up "failing" and killing a squadmate as well. So not only did I lose two squadmates out of a max of three possible, but both of them happened to be the people who were the "correct choices" for the biotic bubble (according to vanilla standards), which probably contributed to my actual pick for the biotic bubble failing and getting yet another squadmate killed. And you know what... I loved it.
It always felt deeply unrealistic (to the point of actively breaking suspension of disbelief) to me that any military operation could fly in as blind as you do through the Omega-4 Relay at the end of the game and get out with a basically guaranteed success rate, simply by ticking some commonsense boxes beforehand (that the game dialogue is pushing you to check off the whole time, no less). Not to mention that real life is generally not so forgiving as to let you off with a slap on the wrist after just one minor mistake... in real life, when s*** starts hitting the fan, inertia very easily takes over, and it actually takes a lot of effort to reverse the bad direction you're going in (which is a mechanic your numbers support, if I am not mistaken... since once someone is killed, fewer squadmates remain to share the risk of subsequent lethality events, therefore shouldn't each death actually slightly raise the probability of subsequent deaths?). By the time I arrived at the Collector Base, even though I could have lost fewer people, my mind was racing and calculating all the ways in which things could have been even worse (for example, if I had lost squadmates who up until that point, had higher readiness scores, or who had specialty bonuses in other fields, or whom my Shepard had developed stronger personal attachments to, etc.), and it really did make me / my Shepard (insofar I played the game as her character rather than as just myself) feel way more invested in the experience.
So in the end, I do see the point of why this mod only allows the player to install two upgrades. I like your numbers, I like the fact that, even in the best-case scenario, one upgrade has a much higher (but not absolute) likelihood of failure than the others (whose likelihood of success is also not absolute), and I think this is a mechanic that REALLY improves the experience of the whole Suicide Mission in general.
Where my objection comes in (and where I differ from other people who have also objected to the two-upgrade limit) is that I think that the problem with this limit lies not so much in the fact that it exists, but in the way you have sold it to the fanbase. I feel like a lot of the pushback you got from players who felt like they were being "punished" for being blocked from installing all upgrades came from the fact that the language you used suggested that you were essentially disallowing them to perform an action that the vanilla game allows them to do (players tend to download mods because they want to REMOVE restrictions that the game places upon them, not add new ones).
Playing your mod (and even before that, just dissecting the numbers you provided) showed me that this was not, in fact, what you were trying to achieve: you are already allowing a substantial chance of passing the check with no deaths, even in the case that the upgrade in question is not installed, which begs the question why you can't pass that off as "we installed this upgrade, but it failed because something went wrong during the installation process that we didn't catch fast enough". In fact, I think it's a more accurate description of what is going on in your mod, and choosing to describe it as "not installing the upgrade" is prejudicing people (like my former self) away from considering your mod on its own merits, and throwing them off from the very real fact that (upgrade not installed) =/= (automatic death) when your mod is installed. (Because that's what it DOES mean, in the conventions of the RPG genre that this game franchise originally belonged to... if you did not even TRY to install the thing that can save a squadmate, it means that when that squadmate dies, that death is YOUR FAULT. And with the way this mod currently presents things, it implies that players are not allowed to TRY. It's not the chance of failure per se that people are reacting badly to, but being blocked from trying at all.)
I mean, think back to a comment made by Paragon Shepard in the first game, "you can't control what other people are going to do, but you can control how you'll react" (to paraphrase, I forget the exact quote). Essentially, I think fans were reacting poorly because by defining it in this way, it seemed to them that you were taking away their ability to react to the circumstances that were thrown at them, when they would have been much more accepting of changes that were made to circumstances outside of Shepard him/herself, because those are circumstances that are outside the player character's control by very definition. Saying that you're still allowed to install all the upgrades (just like in vanilla), but also creating a higher potential for one of them to fail anyway (especially when your mod introduces a slight failure chance for other mods that are successfully upgraded... notice you got barely any complaints for that) would help maintain the illusion of player agency* by preserving their ability to make choices. The random / outside-of-player choice element is then diverted away from the player's own agency / choice / culpability to the Suicide Mission / Omega-4 Relay itself, which is what the player perceives as the proper place for that element to belong to.
*After all, "preserving the illusion of player agency" is fundamentally what even the vanilla Suicide Mission is about... not everything is up to the player because if it were, it would be beyond boring.
Is it a bit of a semantic difference, a "rose by any other name" kind of thing? Absolutely. But that's how these things tend to work, the packaging matters more than we like to acknowledge. I'm just making this suggestion because I hate to see the good parts of this mod go to waste because of what essentially seems to me a miscommunication between the mod author and the broader fanbase.
Should you choose to take me up on this suggestion (and I would understand if you didn't) there are two ways you can go about it:
1) Go with the same upgrade mechanic that fires for the players who installed your mod in a playthrough where all three upgrades are already installed and apply it to all players, BUT ALSO change how the mod presents it (which would just require editing some strings, unless I am mistaken): allow all upgrades but pick one upgrade at random to have the higher chance of failing upon going through the Omega-4 Relay (rather than making it known to the player which one was "uninstalled"). This will have the effect of making the approach towards the Collector Base even more of a black box, which may make the mod more or less appealing to players in general (depending on their personal preferences).
2) Go with the current default mechanic (not allowing the last upgrade to be installed) on an internal level, but on the front-end, make it seem like the upgrade did install (with some added warning that the last upgrade to be installed has a higher chance of failure due to less time available to fine-tune / troubleshoot before launching the mission itself, which will feel less conveniently arbitrary to players than "of ALL the upgrades you can just find in the game, it just so happens that you're forbidden from even trying to install the last of the three most crucial ones, for no specific reason supported by the in-game canon."). If implementable, this will be the option that allows players to retain some level of choice around which updates to prioritize, which may open up further roleplay / headcanoning possibilities** for players (like myself) who like that sort of thing.
On the downside, I see no way to implement either option (or indeed, to preserve the uneven probability of failure mechanic as it currently stands) without affecting the plot flags themselves in the savegame, which may have consequences for savegames imported into the next game. There are a sizeable number of trilogy fans who play Expanded Galaxy Mod for ME3, and that mod gives you extra bonuses according to the upgrades installed on the SR2 for your imported ME2 save... sure, it's nothing you can't fix with a save game editor, but if there DOES exist a way to implement this without affecting saves carried forward into ME3, I think that would be cleaner from a general modding standpoint.
**For example, for my first run on this mod, I gave some thought as to what kinds of upgrades my Shepard would give priority to, according to her personality, and I came up with this list: aggressive / Renegade Shepards who want to get in there and do some major damage, long(ish)-term consequences be damned, will go for armor + guns and forego shields, while paragon-leaning Sheps who have more a traditional "Alliance officer" mentality, and/or where they are actively tactically planning to minimize casualties, will plan on taking more of a boxer's "jab then immediately step back out of the line of fire" tactic, which will mean opting for shields + guns and sacrificing armor. VERY pessimistically-minded Sheps / Sheps with covert forces training specifically (such as infiltrators) might want to go all-in on protection, at the cost of sheer offensive power, so they'll choose armor + shields and pass on the upgrade guns. It was a fun dimension to add to my normal playstyle. Other players may want to metagame and choose the upgrades with a mind to trying to keep certain squadmates safer than others, though that's not something that personally appeals to me.
Feedback (Played) on Risky Suicide Mission: Two-Upgrade Limit (1 comment)
But once I started charting the numbers on a spreadsheet and figuring out how they interact with one another, I started to understand why you might have done it this way, instead of just implementing an elevated probability of death-despite-installation across the board... I'm not confident enough in my grasp of probability theory to say that I have the right idea of how your mod actually crunches the numbers (not to mention the fact that a word-based description simply cannot get precise enough to truly communicate everything going on at a mathematical level). However, trying out multiple possible ways of calculating the numbers as you described them, it suggested to me that creating unevenness, not just a simple random element, in the survival / death checks is what creates half the tension and suspense of your version. And it was that suspense that comes from genuinely not knowing what is going to happen next that I guessed (since this was at the point I hadn't played the mission yet) that was going to create the correct emotional atmosphere around the approach to the Collector Base... a feeling of suspense that is frankly absent from the vanilla SM, if you've been paying even a little bit of attention through even a first playthrough.
My guess turned out to be right. It turned out I was on the unlucky side during the approach, as not only did my not-installed upgrade result in a death, but one of my installed upgrades also ended up "failing" and killing a squadmate as well. So not only did I lose two squadmates out of a max of three possible, but both of them happened to be the people who were the "correct choices" for the biotic bubble (according to vanilla standards), which probably contributed to my actual pick for the biotic bubble failing and getting yet another squadmate killed. And you know what... I loved it.
It always felt deeply unrealistic (to the point of actively breaking suspension of disbelief) to me that any military operation could fly in as blind as you do through the Omega-4 Relay at the end of the game and get out with a basically guaranteed success rate, simply by ticking some commonsense boxes beforehand (that the game dialogue is pushing you to check off the whole time, no less). Not to mention that real life is generally not so forgiving as to let you off with a slap on the wrist after just one minor mistake... in real life, when s*** starts hitting the fan, inertia very easily takes over, and it actually takes a lot of effort to reverse the bad direction you're going in (which is a mechanic your numbers support, if I am not mistaken... since once someone is killed, fewer squadmates remain to share the risk of subsequent lethality events, therefore shouldn't each death actually slightly raise the probability of subsequent deaths?). By the time I arrived at the Collector Base, even though I could have lost fewer people, my mind was racing and calculating all the ways in which things could have been even worse (for example, if I had lost squadmates who up until that point, had higher readiness scores, or who had specialty bonuses in other fields, or whom my Shepard had developed stronger personal attachments to, etc.), and it really did make me / my Shepard (insofar I played the game as her character rather than as just myself) feel way more invested in the experience.
So in the end, I do see the point of why this mod only allows the player to install two upgrades. I like your numbers, I like the fact that, even in the best-case scenario, one upgrade has a much higher (but not absolute) likelihood of failure than the others (whose likelihood of success is also not absolute), and I think this is a mechanic that REALLY improves the experience of the whole Suicide Mission in general.
Where my objection comes in (and where I differ from other people who have also objected to the two-upgrade limit) is that I think that the problem with this limit lies not so much in the fact that it exists, but in the way you have sold it to the fanbase. I feel like a lot of the pushback you got from players who felt like they were being "punished" for being blocked from installing all upgrades came from the fact that the language you used suggested that you were essentially disallowing them to perform an action that the vanilla game allows them to do (players tend to download mods because they want to REMOVE restrictions that the game places upon them, not add new ones).
Playing your mod (and even before that, just dissecting the numbers you provided) showed me that this was not, in fact, what you were trying to achieve: you are already allowing a substantial chance of passing the check with no deaths, even in the case that the upgrade in question is not installed, which begs the question why you can't pass that off as "we installed this upgrade, but it failed because something went wrong during the installation process that we didn't catch fast enough". In fact, I think it's a more accurate description of what is going on in your mod, and choosing to describe it as "not installing the upgrade" is prejudicing people (like my former self) away from considering your mod on its own merits, and throwing them off from the very real fact that (upgrade not installed) =/= (automatic death) when your mod is installed. (Because that's what it DOES mean, in the conventions of the RPG genre that this game franchise originally belonged to... if you did not even TRY to install the thing that can save a squadmate, it means that when that squadmate dies, that death is YOUR FAULT. And with the way this mod currently presents things, it implies that players are not allowed to TRY. It's not the chance of failure per se that people are reacting badly to, but being blocked from trying at all.)
I mean, think back to a comment made by Paragon Shepard in the first game, "you can't control what other people are going to do, but you can control how you'll react" (to paraphrase, I forget the exact quote). Essentially, I think fans were reacting poorly because by defining it in this way, it seemed to them that you were taking away their ability to react to the circumstances that were thrown at them, when they would have been much more accepting of changes that were made to circumstances outside of Shepard him/herself, because those are circumstances that are outside the player character's control by very definition. Saying that you're still allowed to install all the upgrades (just like in vanilla), but also creating a higher potential for one of them to fail anyway (especially when your mod introduces a slight failure chance for other mods that are successfully upgraded... notice you got barely any complaints for that) would help maintain the illusion of player agency* by preserving their ability to make choices. The random / outside-of-player choice element is then diverted away from the player's own agency / choice / culpability to the Suicide Mission / Omega-4 Relay itself, which is what the player perceives as the proper place for that element to belong to.
*After all, "preserving the illusion of player agency" is fundamentally what even the vanilla Suicide Mission is about... not everything is up to the player because if it were, it would be beyond boring.
Is it a bit of a semantic difference, a "rose by any other name" kind of thing? Absolutely. But that's how these things tend to work, the packaging matters more than we like to acknowledge. I'm just making this suggestion because I hate to see the good parts of this mod go to waste because of what essentially seems to me a miscommunication between the mod author and the broader fanbase.
Should you choose to take me up on this suggestion (and I would understand if you didn't) there are two ways you can go about it:
1) Go with the same upgrade mechanic that fires for the players who installed your mod in a playthrough where all three upgrades are already installed and apply it to all players, BUT ALSO change how the mod presents it (which would just require editing some strings, unless I am mistaken): allow all upgrades but pick one upgrade at random to have the higher chance of failing upon going through the Omega-4 Relay (rather than making it known to the player which one was "uninstalled"). This will have the effect of making the approach towards the Collector Base even more of a black box, which may make the mod more or less appealing to players in general (depending on their personal preferences).
2) Go with the current default mechanic (not allowing the last upgrade to be installed) on an internal level, but on the front-end, make it seem like the upgrade did install (with some added warning that the last upgrade to be installed has a higher chance of failure due to less time available to fine-tune / troubleshoot before launching the mission itself, which will feel less conveniently arbitrary to players than "of ALL the upgrades you can just find in the game, it just so happens that you're forbidden from even trying to install the last of the three most crucial ones, for no specific reason supported by the in-game canon."). If implementable, this will be the option that allows players to retain some level of choice around which updates to prioritize, which may open up further roleplay / headcanoning possibilities** for players (like myself) who like that sort of thing.
On the downside, I see no way to implement either option (or indeed, to preserve the uneven probability of failure mechanic as it currently stands) without affecting the plot flags themselves in the savegame, which may have consequences for savegames imported into the next game. There are a sizeable number of trilogy fans who play Expanded Galaxy Mod for ME3, and that mod gives you extra bonuses according to the upgrades installed on the SR2 for your imported ME2 save... sure, it's nothing you can't fix with a save game editor, but if there DOES exist a way to implement this without affecting saves carried forward into ME3, I think that would be cleaner from a general modding standpoint.
**For example, for my first run on this mod, I gave some thought as to what kinds of upgrades my Shepard would give priority to, according to her personality, and I came up with this list: aggressive / Renegade Shepards who want to get in there and do some major damage, long(ish)-term consequences be damned, will go for armor + guns and forego shields, while paragon-leaning Sheps who have more a traditional "Alliance officer" mentality, and/or where they are actively tactically planning to minimize casualties, will plan on taking more of a boxer's "jab then immediately step back out of the line of fire" tactic, which will mean opting for shields + guns and sacrificing armor. VERY pessimistically-minded Sheps / Sheps with covert forces training specifically (such as infiltrators) might want to go all-in on protection, at the cost of sheer offensive power, so they'll choose armor + shields and pass on the upgrade guns. It was a fun dimension to add to my normal playstyle. Other players may want to metagame and choose the upgrades with a mind to trying to keep certain squadmates safer than others, though that's not something that personally appeals to me.