I understand your concerns and that was why I suggested an option at setup to use this or not. However I do believe what you characterize as a player getting randomly screwed is actually how things go IRL. Today I just happened to be browsing through a list of failed aircraft, both civilian and military that were designed and built with high hopes and just didn't work out as planned.
Something that we are all probably familiar with today that speaks to this is all of the vaccines that are under development some of which are now out for use but some of which are failed and no longer being pursued. I don't see this as the player being randomly screwed I just see this as more of a situation that exists where you don't know exactly what the return will be on the research path you choose and you will either live with pushing down a certain line of research even though there may be a better one or more broadly pushing several paths to get a feel for which might be the best way to go this particular time.
If research and development was as cut and dried as we find it in games than every country would field basically the same tanks, planes and ships. This is not the case because real research has a huge degree of uncertainty and results are often not what was anticipated when the project was begun.
One of the tricky things about a game, even a "realistic" one, is that real life itself actually makes a terrible game. This is because a game is at the end of the day meant to be fun, while real life has no such developer's mandate and no such need to sell copies/pad download numbers. I don't believe that for most players, failing at random would be considered "fun" - if anything, random elements in games work best when they make a game unpredictable and present challenges for a player to overcome -
Precursors are an excellent example of this.
As a player however, Aurora gives you considerable leeway for roleplay and that means you can have the freedom and flexibility to play in such a way that you get a mix of successful and failed component and ship designs if you choose to play more experimentally instead of seeking the purely optimal designs. Sure, you can just build every ship with 50% engine mass and railguns, or lock onto a whole-hog carrier + missile bomber doctrine which is proven to be highly effective, but where's the fun in that? (Rhetorical question; for many people this is fun and they are very happy with it. Weirdos...)
There are really quite a lot of variables in Aurora to design with, so certainly it's possible to design and build many "failed" designs in the process of finding a good balance of different components or even the doctrine of the designs. In my current campaign I ran into an aggressive NPR very early and had to wage a tense defense with my starting ship designs and insufficient logistics to support them. Notably my mixed torpedo bomber / Gauss fighters ran into numerous problems due to a combination of superior enemy tech (the NPR invested a lot of RP into shields) and, uh, poor naval doctrine (also known as PEBKAC) and were ineffective. Because of this, my next generation of carrier fighters will be designed very differently, as my initial designs did not work out as well as planned or hoped for (much like, incidentally, the IRL British carrier aircraft they were based on).
The difference here is that as a player of a game, I can recognize where my plans failed and devise an approach to correct those failures. The research system in Aurora lacks that capability in the sense that it exists in real life. In reality, if a new technology or vehicle doesn't work out, the researchers learn from how that technology performed and figure out what needs to be fixed in the next designs. In Aurora we don't have that aspect in the research mechanic itself, and adding a random modifier doesn't give that either - if my tech fails, I can't learn anything from it, all I can do is roll the dice again (for the low, low price of another 5,000 RP).
