To Boldly Go... (a Starfleet quest)

[SuperSonicSound]

Just to verify, is the odds of a catastrophic failure, one that leaves some or all of the crew dead, require hitting that 6% number twice in a row?

Admittedly, even the first type of failure in the middle of a 'save the planet/sister ship by SCIENCE!' mission, sounds scary, but I don't think those sort of missions would be the ones this type of ship would run into without backup.

...right?

Click to shrink...

If you have a 6% chance of failure you have a .6% chance of a severe problem, if the warp core and hull are high reliability then the craft has basically no chance of being destroyed.

40% of all reliability issues are 'funny mishaps' which could generate RP. 10% are severe and 50% are problems. So i personally find a craft of 94-95% reliability fine as long as the warp core and hull arent impacted.

 

[AlphaDelta]

It doesn't change much in the sense that the rest of your analysis was completely meaningless anyway. There is no sensible reason to aggregate failure chance like that for science ships.
That would only make sense if a single failure anywhere was disastrous for the whole group. That might make sense for a combat fleet, but not for a collection of independently operating science ships. What you are doing is the equivalent of saying that enough people bathing makes the risk of bathing tubs unacceptable because it's virtually guaranteed someone slips and breaks their ne

Click to shrink...

There's a .94 chance of failure. Now, to get the chance for a failure across the fleet, you multiply the individual chances.

....I can't believe I'm lecturing people on probability on the internet (askreddit/eli5/stackexchange excluded), given I dislike combinatorics.

I was active in that thread and recognize no such consensus. Even if there was a narrow group forming such a consensus there would have no right to impose it here.

Click to shrink...

I'll concede that not everyone agrees, but all of the current designs use .995 as the gold standard.

 

[Briefvoice]

Except we don't get to choose what missions they go on(it's rolled randomly to see who's in range to respond), so we need to be prepared for many different types of missions

Click to shrink...

Though this gets safer the more ships you have. The example posted shows:
1. Ships are more likely to make the roll if it's in their specialty (Defense + Science versus Defense + Presence or presumably Defense + Combat for some mission).
2. If multiple ships make the roll, the QM may choose the ship that's actually the most appropriate to respond.

So it's not totally random.

Sorry. The actual number is ~50% for 10 ships each year. A rough estimate for bad events vs. funny would be 30%, which is unacceptably high.

All of my figures are for 1 ship; there's the chance that multiple could get hit by breakdowns.

Click to shrink...

I think there's something everyone is missing, though. A failed reliability roll means that the ship suffers an event that will create damage if the Event difficulty roll isn't passed, right?

What is that roll likely to be based on? That's right, SCIENCE. A dedicated science ship doesn't care if it's unreliable, because if the hull starts falling apart the crew will just science the shit out of it with their massive bonus and come through the Event okay.

 

[SuperSonicSound]

There's a .94 chance of failure. Now, to get the chance for a failure across the fleet, you multiply the individual chances.

....I can't believe I'm lecturing people on probability on the internet (askreddit/eli5/stackexchange excluded), given I dislike combinatorics.

I'll concede that not everyone agrees, but all of the current designs use .995 as the gold standard.

Click to shrink...

No he is right your logic doesnt hold up correctly because a ship being out of service in the Sol sector has no impact on any other ship, all your calculation is determining is the number star base berths we might need.

 

[AlphaDelta]

Though this gets safer the more ships you have. The example posted shows:
1. Ships are more likely to make the roll if it's in their specialty (Defense + Science versus Defense + Presence or presumably Defense + Combat for some mission).
2. If multiple ships make the roll, the QM may choose the ship that's actually the most appropriate to respond.

So it's not totally random.



I think there's something everyone is missing, though. A failed reliability roll means that the ship suffers an event that will create damage if the Event difficulty roll isn't passed, right?

What is that roll likely to be based on? That's right, SCIENCE. A dedicated science ship doesn't care if it's unreliable, because if the hull starts falling apart the crew will just science the shit out of it with their massive bonus and come throguh the Event okay.

Click to shrink...

It's a breakdown, so presumably it might be something like "navigational deflector breaks, now you can't travel at speed, because debris"

 

[Briefvoice]

It's a breakdown, so presumably it might be something like "navigational deflector breaks, now you can't travel at speed, because debris"

Click to shrink...

No, it's an event like any other. It starts a plot thingie happening, just like if an event was rolled in the sector. It's not a guarantee of anything bad happening any more than responding to a distress call is.

 

[Yeangst]

There's a .94 chance of failure. Now, to get the chance for a failure across the fleet, you multiply the individual chances.

....I can't believe I'm lecturing people on probability on the internet (askreddit/eli5/stackexchange excluded), given I dislike combinatorics.

I'll concede that not everyone agrees, but all of the current designs use .995 as the gold standard.

Click to shrink...

He's not arguing against the math ( though it's actually 1-prob(no failures)) but that's relevant. He's saying that the relevant stat is individual failure, not anything aggregated.

I'm inclined to agree. Science ship, until we get to Intrepids, are probably going to be meant for within borders type missions. I'm fine if every few years we have to send one or two back to a shipyard.

 

[Sightsear]

I'll concede that not everyone agrees, but all of the current designs use .995 as the gold standard

Click to shrink...

Would we, of the low standards faction, forming a mob and marching over there, demanding higher stat ships in exchange for marginally lower reliability, change this attitude at all?

Mostly joking. But seriously, you guys need to loosen up.

 

[Nix]

There's a .94 chance of failure. Now, to get the chance for a failure across the fleet, you multiply the individual chances.

....I can't believe I'm lecturing people on probability on the internet (askreddit/eli5/stackexchange excluded), given I dislike combinatorics.

Click to shrink...

Get off your high horse. You know you aren't the only one who has ever taken math classes, right? To "lecture" someone you have to tell them something they don't already know.

Your calculation would (at a first glace) have been correct (except turn vs year), but completely meaningless. The chance "for a failure across the fleet" is not in the least bit relevant. Combinatorics is simply the wrong tool for the problem. Expected value is what matters here.

 

[AKuz]

Why is so much math happening?

Angry math everywhere ;_;

 

[AlphaDelta]

Get off your high horse. You know you aren't the only one who has ever taken math classes, right? To "lecture" someone you have to tell them something they don't already know.

Your calculation would (at a first glace) have been correct (except turn vs year), but completely meaningless. The chance "for a failure across the fleet" is not in the least bit relevant. Combinatorics is simply the wrong tool for the problem. Expected value is what matters here.

Click to shrink...

Sorry if I'm sounding frustrated.

though it's actually 1-prob(no failures)

Click to shrink...

Ok. In that case, we have:
P(No failures on one ship) = .94
P(No failures on n ships) = (.94)^n
P(!(No failures on n ships)) = 1 - (.94)^n

For n = 10, we have P = .46.

RL warships are not designed to the standard of "There's a 95% chance everything will be fine".

No, it's an event like any other. It starts a plot thingie happening, just like if an event was rolled in the sector. It's not a guarantee of anything bad happening any more than responding to a distress call is.

Click to shrink...

The term "breakdown" gives the impression of something going wrong on the ship.

No he is right your logic doesnt hold up correctly because a ship being out of service in the Sol sector has no impact on any other ship, all your calculation is determining is the number star base berths we might need.

Click to shrink...

I'm estimating probability of a ship out of the fleet experiencing a damaging breakdown. It's unsustainable, and we don't need "breakdowns of the week".

I have a theory as to why Voyager had so many breakdowns of the week – they had to use too many fudge factors to fit the Galaxy's equipment in.

Why is so much math happening?

Angry math everywhere ;_;

Click to shrink...

Fudge factor debate:
99.5 vs 94%

 

[SuperSonicSound]

If you want to work out something meaningful the hypothetical is having two-three of this craft in every major sector, what is the chance more than one in a given sector is knocked out of operation on a given quarter.

Presuming 95% reliability. As long as two or more remain operational the increased stats in R and science out weigh the extra vessel that could be present with higher reliability.

Aggregating failure chance across the whole galaxy doesnt provide meaningful numbers because the impact of any one ship having some kind of break down is localised, on top of that the numbers need to be refined a step further because nearly half of the break down chance results in something completely trivial.


Also keep in mind that unless something severe happens our starbases not our berths repair escort sized vessels.

 

[Yeangst]

I don't we should have a flat reliability minimum across our fleets.

Yes, it's vital for Explorers, since they often operate far from any support, represent the Federation diplomatically, and are the linchpins to any crisis response.

Cruisers, we can afford to take a few more risk with, but not much. They are the key to defense.

If an escort or two breaks down, it's not a huge deal. We should have a lot of them, so the chance of at least one breaking down may sound scary, but we can manage. Maybe we should a different metric like expected amount of combat power sidelined.

Science ships, like I said before, we can take a few risks with.

 

[AlphaDelta]

If you want to work out something meaningful the hypothetical is having two-three of this craft in every major sector, what is the chance more than one in a given sector is knocked out of operation on a given quarter.

Presuming 95% reliability. As long as two or more remain operational the increased stats in R and science out weigh the extra vessel that could be present with higher reliability.

Click to shrink...

We don't have enough to afford the loss.

We don't have enough to garrison 2-3 in each sector. Right now, we have 10 sectors. We're probably going to have at least 15 by the time we have 10 of these. To garrison our existing home sectors plus BZs with 2-3, we'd need 14-21. And that doesn't count the new sectors we will get (Caitians, etc.)

Plus, research time is a thing.

I don't we should have a flat reliability minimum across our fleets.

Yes, it's vital for Explorers, since they often operate far from any support, represent the Federation diplomatically, and are the linchpins to any crisis response.

Cruisers, we can afford to take a few more risk with, but not much. They are the key to defense.

If an escort or two breaks down, it's not a huge deal. We should have a lot of them, so the chance of at least one breaking down may sound scary, but we can manage. Maybe we should a different metric like expected amount of combat power sidelined.

Science ships, like I said before, we can take a few risks with.

Click to shrink...

We don't have enough to afford losing a couple.


I now understand why TNG Federation had so many old ships and kitbashes – there were too many sectors to staff with current-gen and previous-gen ships

 

[Sightsear]

Why is so much math happening?

Angry math everywhere ;_;

Click to shrink...

Sorry, but we're at a phase where we need to determine whats more important:
Modestly powered, but nigh-perfect reliability ships.
Or overpowered, almost perfectly reliable ships.

Seriously, dropping our reliability standards by a few percent points, and you can fit some goodies in there.

 

[Void Stalker]

But for Oberth and its replacement we are likely to only have one per sector since they require a tech heavy crew, if that goes away we lose coverage in a sector. What I think will help is if we can see a coup of simulated reliability rolls and also see what happens on rolls where it fails the initial reliability so that something happens. Getting a peak at how ships respond has given us a much better idea of how to allocate sector fleets and what ships we need for response.

 

[Yeangst]

Ok. In that case, we have:
P(No failures on one ship) = .94
P(No failures on n ships) = (.94)^n
P(!(No failures on n ships)) = 1 - (.94)^n

For n = 10, we have P = .46.

RL warships are not designed to the standard of "There's a 95% chance everything will be fine".

Click to shrink...

Yes they are. Aircraft Carriers spend more than half of every year in dock, undergoing maintenance. Military aircraft need something like 20 man-hours of maintenance for every hour of flight. Commercial aircraft aren't much better.

Nothing in real life is designed for 99.5% chance nothing happens in a year. Not even a Toyota Camry.

We don't have enough to afford losing a couple.

Click to shrink...

That's problem with using the probability of not getting a single failure in your entire fleet is this: When we can't afford to lose a couple to random breakdowns, great. We probably won't because n is low. When n is high, the chance of having a breakdown or two increases, but we'll be fine because n is high.

 

[SuperSonicSound]

We don't have enough to afford the loss.

We don't have enough to garrison 2-3 in each sector. Right now, we have 10 sectors. We're probably going to have at least 15 by the time we have 10 of these. To garrison our existing home sectors plus BZs with 2-3, we'd need 14-21. And that doesn't count the new sectors we will get (Caitians, etc.)

Plus, research time is a thing.

We don't have enough to afford losing a couple.

Click to shrink...

Enough what? Resources to build them? As escorts they would be rather cheap, is it berths in case of failure? Starbases can repair escort sized vessels we wouldnt need to use ship yard berths for any ship that goes down unless its a severe failure which on 6% reliability ship is going to be .6% overal.

Even with twenty of these things we'd be looking at most one or two suffering any kind of problem at all across every sector. Which the sector starbase would end up dealing with the majority of the time.

 

[Nix]

Also, it would be a shame if a potentially fun mechanic that was inspired by a hilarious fan theory never saw proper use. I agree with wanting our Explorers and defence cruisers to be reliable, but the reasons for that for the most part don't apply to science ships. Oneiros would get to write fun log entries, most of us would have a laugh reading them, and we'd very likely earn more rp on average. Ace a pulsar flyby due to high stats, miss a proto-planet collision the next month, but get bonus progress to warp field stabilization. That sort of thing.

 

[AlphaDelta]

Yes they are. Aircraft Carriers spend more than half of every year in dock, undergoing maintenance. Military aircraft need something like 20 man-hours of maintenance for every hour of flight. Commercial aircraft aren't much better.

Nothing in real life is designed for 99.5% chance nothing happens in a year. Not even a Toyota Camry.
That's problem with using the probability of not getting a single failure in your entire fleet is this: When we can't afford to lose a couple to random breakdowns, great. We probably won't because n is low. When n is high, the chance of having a breakdown or two increases, but we'll be fine because n is high.

Click to shrink...

We won't be fine.

To garrison our existing deployment areas, we need 8 Keplers. There's a 40% chance of one of them getting taken out, leaving that sector without a Kepler.

Enough what? Resources to build them? As escorts they would be rather cheap, is it berths in case of failure? Starbases can repair escort sized vessels we wouldnt need to use ship yard berths for any ship that goes down unless its a severe failure which on 6% reliability ship is going to be .6% overal.

Even with twenty of these things we'd be looking at most one or two suffering any kind of problem at all across every sector. Which the sector starbase would end up dealing with the majority of the time.

Click to shrink...

My concern is with redundancy.

 

[AlphaDelta]

Also, it would be a shame if a potentially fun mechanic that was inspired by a hilarious fan theory never saw proper use. I agree with wanting our Explorers and defence cruisers to be reliable, but the reasons for that for the most part don't apply to science ships. Oneiros would get to write fun log entries, most of us would have a laugh reading them, and we'd very likely earn more rp on average. Ace a pulsar flyby due to high stats, miss a proto-planet collision the next month, but get bonus progress to warp field stabilization. That sort of thing.

Click to shrink...

That's the normal events.

We already destroyed a moon, and got bonus rp.

These breakdowns have a higher chance of bad things happening. If good events were in the majority, I'd agree.

FYI, with all 5s, the Kepler can go up to 11 at the cost of 80% reliability.

 

[Yeangst]

We won't be fine.

To garrison our existing deployment areas, we need 8 Keplers. There's a 40% chance of one of them getting taken out, leaving that sector without a Kepler.

Click to shrink...

Right now, none of our sector's have Keplers.

 

[SuperSonicSound]

We won't be fine.

To garrison our existing deployment areas, we need 8 Keplers. There's a 40% chance of one of them getting taken out, leaving that sector without a Kepler.



My concern is with redundancy.

Click to shrink...

Sure if you're only putting one in each sector but given the nature of the world our ships operate in we want two or three in every sector. One science vessel is kind of pointless because with just one the failure chances on not meeting reaction rolls will have more impact than the ships reliability ever could.

Your problem is you are running reliability effects in isolation of everything else.

 

[Sightsear]

I'm honestly debating nagging Oneiros to put up a poll to see what the silent majority thinks... Seems like overkill at the moment though, although there is a bit of a closed circuit among our engineers. Might be nice to shake them out of their complacency.

 

[AlphaDelta]

I'm going to take a break from this – I'm tired of arguing.