|
stilicho410 -> RE: Some thoughts on reliability (1/17/2014 5:35:53 PM)
|
This turned out to be a very long post, so I'll summarize my suggestions right at the start.
1. The "diminishing returns" R&D model could be fine-tuned a bit to be slightly less asymptotic.
2. The reliability gains from successful mission should be increased to what they were in BARIS and the first few BASPM builds, perhaps 1-2%.
3. Revert to the BARIS system of only inflicting reliability penalties when people are killed. Otherwise, if any steps succeed for a component, apply a reliability gain.
I'm sure there are a variety of opinions, but I find the "diminishing returns" formula for R&D very dispiriting, especially for the later model rockets.
I started researching the Saturn 1B in Q1.1968 with 4 researchers with Skills ranging from 96 to 93. I didn't reach Max R&D reliability until Q4.1971.
I started researching the Saturn V around Q1.1969 with 4 researchers with Skills initially ranging from 93 to 90, later 96 to 93.
At Q4.1971, I was still 5.4% away from Max R&D reliability, and the increments were down to 1%.
The Saturn V finally reached Max R&D reliability in Q1.1973, and that was with the help of 6 successfull missions I launched after it got to 89%.
I realize that historical development of the Saturn rockets lasted many years, but historically they started around 1960 with the first unmanned Saturn 1B test in Feb 1966; with the current game mechanics there is no way a player would be able to start that soon.
Back around version 0.74, I thought perhaps the diminishing R&D returns was to make the players consider launching unmanned tests before Max R&D, as reliability increases for successful missions in 0.74 generally ran a couple percent. That made some sense, there comes a point where you'd gain more from an actual test rather than more R&D. I think that reflected history fairly well; by my count, there were just 2 unmanned Gemini/Titan launches, 2 unmanned Apollo CSM/Saturn 1B launches, and 2 unmanned Apollo CSM/Saturn V launches.
However, the last several versions have shifted to giving a fraction of a percent for successful missions, too. The current reliability model is starting to look like an asymptotic curve, where "Max Reliability" will never be reached.
BARIS gave a flat 1% improvement for successful missions. BASPM could be more flexible, giving maybe 1% to earlier spacecraft/boosters like the Mercury program, and 2% for later generations like Gemini and Apollo. I think that would give a more historical rate of progression. I think it would also improve gamer enjoyment; After running the risk of a launch, I found it very dispiriting to only get a fraction of a percent.
I already mentioned this upthread, but I still submit that for Reliability decreases, the BARIS model of only inflicting them when people got killed, and giving a reliability gain when any mission steps succeeded, worked better both historically and for gameplay. Here are some additional thoughts:
1. Currently when launching a unmanned test mission, one is balancing an infinitesimal gain vs a 20% penalty, which under the diminishing returns R&D model will take literally years to make up. The only incentive is to avoid the reliability penalty of a skipped mission configuration, so there's no reason to do more than one unmanned test of any given type.
2. Historically, they often launched at least a couple of any given type of unmanned probe for redunancy. (Pioneer 10 and 11; Viking 1 and 2, etc). Currently, there is no gain from doing this within BASPM, if you send 2 Pioneers to Jupiter and 1 fails, the remaining will get a big reliability hit and likely fail, too.
|
|
|
|
