Slitherine

Hello,

So, even now that the % Mission Step Success Probability has been properly renamed, there is something odd. Or I'm the luckiest person on Earth at this game. I just had two perfect X-15 missions, without failures. And this is highly surprising, given the % success rate of the gauge.

Let's be generous and say it was 80% (it was a bit less in my game). There are 7 steps in a X-15 mission. Statistical mathematics say to us, P beign the overall mission success probability :

Mission 1 :

P1 = 0.8 * 0.8 * 0.8 * 0.8 * 0.8 * 0.8 * 0.8
P1 = 0.2097152

...which means I should have roughly 21% chances to have a fully successful mission. 1 on 5 if you prefer. That's not a lot.

Mission 2 :

This time we start from 0.21, I round up the previous result.

P2 = 0.21 * 0.8 * 0.8 * 0.8 * 0.8 * 0.8 * 0.8 * 0.8
P2 = 0.044040192

...which means I should have roughly 0.4% chances to have a second fully successful mission in a row. That's getting very improbable.

So I'd like to know what I am missing, given that BARIS was working that way, and that you had to get 95%+ safe hardware and astronauts bonus points to have a serious hope on a complex mission, like a lunar landing.

Edit : OK, 3 in a row now. And my gauge wasn't displaying 80%, but 68%. I let you redo the maths, this time we are clearly under 0.01% ...

http://peyre.x10.mx/BARIS/BARISTips_v2.txt

Read the "Reliability ..." part - it should be pretty much spot on for BASPM, as well.
This is the probability that something goes wrong, not that your mission outright fails. With those odds you calculated, nobody would've ever made it to the moon in BARIS.

OK so right now, only catastrophic failures are taken into account ?

In this case, I guess that minor failures will allow the mission to carry one but will make an abort highly recommended and add a serious prestige penalty at the end, depending of the gravity of what happened. Right now it's "all or nothing".

N_Molson wrote:OK so right now, only catastrophic failures are taken into account ?

We'd need Ignacio for an answer to that. Would like to hear it, too.
Some kind of feedback would be nice, and if it's just to see that minor problems do occur and are handled by the correct entity in-game.
I've had several X15 missions fail, 2 of them "in a row" - after researching the damned thing up to 80% again, and with reasonably good personnel. So, I appear to have the rotten luck with it you are missing.

Well, this is getting weird. Now I went through Mercury and I just completed suborbital Gemini without a single failure.

Hello all,

Some clarifications about this: at the moment the 'reliability' for a mission component is not exactly that. What it means is more like "the chance of this component of passing all the mission steps it's involved for the current mission", which is not the best solution since the value remains the same irrespective of whether the component is used for a mission with 10 steps or a mission with 100.

I'm planning to rework all that for the next update, so that the 'reliability' becomes the true reliability. I didn't want to introduce this change in 0.7.4 because it required substantial changes, not only to the code but also to the data, and I wasn't going to be able to test it properly. Bare with me, fingers-crossed I'll be able to implement it during the weekend and you'll get it on 0.7.5 next Friday. The idea is that the max RD values will go up considerably, pretty much over 99%. Otherwise the missions will become impossible to pass.

Cheers,

I think the way the game is made to work is that sometimes you will have failures but your crew and or flight controllers will fix the problem and the mission will continue.

Ah, sorry, yes. I neglected to explain the involvement of flight controllers and the crew!

At the moment, flight controllers only kick-in if there's been a failure with the mission components. Think of them as a 'saving throw' from Dungeons & Dragons. The game would multiply their skills by the reliability of the mission components. Highly skilled controllers will leave the components reliability untouched whereas less competent ones will significantly reduce it. Then it throws a dice between 0 and 100 and, if the result if greater than the product of the skill and the components reliability, the step fails.

Flight crews use a similar algorithm, but they are not applied as a 'saving throw'. Basically, all the relevant skills for crew members for a particular step are averaged together, then divided by 100 in order to get a value between 0 and 1. That value then gets multiplied by the components reliability. Notice that I'm planning to change the algorithm for flight crews involvement, because even a highly qualified 90% individual would reduce the reliability of a 99.5% to 89.55%, and that would make the game very difficult to play.

By the way, feel free to let me know if you want to introduce changes to the current system.

Cheers,

Ah, thanks for the clarifications. Yes, I was thinking that the max. R&D & safety values were quite low, though that should vary between the programs.