I'm always appreciative of actual experiments with controls and numbers instead of anecdotes! You did these tests in sector, but it would be interesting to see the difference in performance with out of sector ships. I've found the rate of mining to be vastly increased with higher overall ship skill level when observing ships out of sector. Their maneuverability is not affected, they simply get more minerals per unit of time with higher skill.
Thank you! Gathering data takes a lot of time which is probably why everyone uses anecdotal evidence instead 😁 . These tests were done OOS, I was tucked away in a corner at Nopileos 2 so to not disrupt SETA. The background footage was filmed afterwards. High skilled pilots will yield more minerals but I kept the pilot at 2 stars as I feel that most players use their higher skilled pilots for autotraders, carriers, and destroyers. That way the results will be similar to what an average player should expect.
@@JKNinja The crew has a moderate effect of overall ship skill level, but the pilot by far has the biggest influence. It is overall ship skill level that determines mining efficiency. I did my own tests where I had several ship mining a single patch at once, and recorded the overall skill level of each ship, plotted it with skill level on the X axis and observed mining rate on the y axis, and saw an almost exponential and very significant increase in mining speed as crew level increased.
I did a T-test of the results from the miner data. There is a 1% chance that the improvement seen is due to randomness in the medium miners, and 4% for the large miners. I didn't bother with the small miners as the improvement seen was even larger than the mediums. By convention in statistics these values are considered statistically significant. It's reasonable to draw the conclusion that the crew does impact performance. There are online T-test calculators that are pretty easy to use, you just type in your data sets and they do the calculations. It's nice knowing what the chances are for an improvement to be random chance. Of course if you want to explain the math behind it, it's a bit of a deep dive.
Threw the results into a spreadsheet and discovered something interesting that could be worth further investigation. There may be more to crewing than just a bump in average productivity. The standard deviation of the runs for the small ships more than doubled for the crewed runs compared to uncrewed. For those that don't know what that means, standard deviation is a way of measuring the spread of a set of numbers relative to the average. A higher standard deviation means the difference between the highest and lowest numbers is going to be greater. For it to more than double for the small ships means a crewed small ship will have much larger variation between runs than an uncrewed small ship. So crews may only be worth it if you plan to do multiple runs. On the other hand, the medium ships cut their standard deviation in half by adding a crew. So not only was the average better, their performance was substantially more consistent. The large ships had their standard deviation increased by crewing as well. But only by about 22%. Would specialists have made a difference in that? I don't know. Unfortunately, it's a rather small sample set. So there's no telling if those numbers mean anything or if they're just an artifact.
For the trader test you could limit how far they are allowed to jump to one sector and put them in nopileos fortune, that would mitigate a lot of variance while experimenting with trade over distance so that travel time remains a factor.
an idea for handling the variance, A) large sample size, probably time consuming, but should eat most of the variance B) fixed or semi-fixed routes, restricting the range to only a few sectors, as I suspect your biggest issue is gonna be gate weirdness, and if crew improve maneuverability, then it's likely they will reduce the weirdness using gates and highways. C) actually calculating the 95% confidence interval, easy to do in excel, and will tell you if you can exclude no effect, since if the 95% confidence interval includes 0, we don't know if any effect exists
I kind of just throw service crew everywhere eventually. Most games I'll run two Heracles Vanguards that constantly get fed with rookies and before I know it I have 420 5-star engineers. When it comes to profits-per-hour I could just merc a bunch of Honshu for a quick 30-40 mil per heist if things are moving too slowly for my desired rate of expansion.
i think the best way to test traders, would be to have them on adv autotrade, anchor them to a sector with a shipyard and wharf, and have them only trade ship building components. let them travel a maximum of 1 sector. this way they wont be flying all over the place. they will be anchored to a small area.
Doesnt this mean that maneuverability is super important even for economy? Doesnt that mean that investing in the Mk3 combat thrusters will ALSO bring a similar improvement in profit per hour? Even if crewed by amateur crewmen, the increased maneuverability (pitch yaw etc) would increase profits?
You can boost while charging travel mode, which the AI will not do and instead wait out the three seconds. Combined with AI pathing that will fly them into asteroids and reset travel speed it can sometimes be slower in an asteroid heavy sector. The AI has an easier time escaping from Kha'ak and Xenon with all-around drives, which didn't come into play in this video, but is also something to consider
@JK Ninja new player here would love to see this test on all around vs travel if you have one ive been using travel because i figured AI would be slowest going back and forth as you say because they randomly fly across the galaxy sometimes for the asteroid lol
Sweet! this is very helpful, thank you! Now... I'll feel even worse with a ship full of dozens of people explodes because Kha'ak are a thing. ALSO. side note, not entirely related to the video. I like how you pronounce Kha'ak and while I get a school boyish chuckle from saying it like "cock" all the time - I think pronouncing it more like Kayak (not exactly how you pronounced it, but close) - is funnier to me.
I always tend to use 80% crew 20% marines for larger more expensive ships and 100% crew for smaller less expensive. I did notice on M sized ships with not much crew that inside of them crew members would jump between stations as things happened. Sometimes it didnt effect anything like when traveling and docking, but in combat things seemed more hectic. Of course thats the AI in general when you're in system during combat.
The key difference is in equipment mod compare 10M to 1L for similar % improvememt and u use almost same mods for both (bar turrets). T3 mods on your L miners beats out M miners in overall cost effectiveness IMO. U want drones on Hull.
service crew will repair damaged hulls, I believe they will only repair it up to 15-75% depending on engineer skill. On large ships they will also repair damaged modules such as turrets and engines. Is there a particular combat scenario(s) you are interested in testing? If I can get enough of those to make a video I'd be happy to make one!
I've only been playing a few months and noticed that turrets are apparently automtomus. I've flow ships which I accidently didn't have any crew meaning service or marines and had my turrets set on defend. When doing escort mission and came under attack the turrets started defending. Thus I'd assume that service are not as important to combat and marines are only valuable if and when boarding my ship or when I send them to board another.
The AI will come to a full stop before reengaging the travel drives instead of using fight assist off to maintain momentum. If they need to make a hard turn or flying through asteroids it will severely slow them down
.........1:19 i see default response to pirates....... ughh....... 100 hours playing i never seen that. they have been the bain of my existance. i loose so many ships to them cause i miss the msg and so my cargo crew says well boss said nothing so i guess we atk them them......... or i have to chase them down to save them time after time......... i wish i seen that 90 hours ago lol
The turrets won't shoot faster. Higher level service crew will repair the hull to a higher percentage while repair drones are better for damaged modules.
I love how the ship runs straight into the station in the background as you talk about how good they did.
I'm always appreciative of actual experiments with controls and numbers instead of anecdotes! You did these tests in sector, but it would be interesting to see the difference in performance with out of sector ships. I've found the rate of mining to be vastly increased with higher overall ship skill level when observing ships out of sector. Their maneuverability is not affected, they simply get more minerals per unit of time with higher skill.
Thank you! Gathering data takes a lot of time which is probably why everyone uses anecdotal evidence instead 😁 .
These tests were done OOS, I was tucked away in a corner at Nopileos 2 so to not disrupt SETA. The background footage was filmed afterwards. High skilled pilots will yield more minerals but I kept the pilot at 2 stars as I feel that most players use their higher skilled pilots for autotraders, carriers, and destroyers. That way the results will be similar to what an average player should expect.
@@JKNinja The crew has a moderate effect of overall ship skill level, but the pilot by far has the biggest influence. It is overall ship skill level that determines mining efficiency.
I did my own tests where I had several ship mining a single patch at once, and recorded the overall skill level of each ship, plotted it with skill level on the X axis and observed mining rate on the y axis, and saw an almost exponential and very significant increase in mining speed as crew level increased.
I did a T-test of the results from the miner data. There is a 1% chance that the improvement seen is due to randomness in the medium miners, and 4% for the large miners. I didn't bother with the small miners as the improvement seen was even larger than the mediums. By convention in statistics these values are considered statistically significant. It's reasonable to draw the conclusion that the crew does impact performance.
There are online T-test calculators that are pretty easy to use, you just type in your data sets and they do the calculations. It's nice knowing what the chances are for an improvement to be random chance. Of course if you want to explain the math behind it, it's a bit of a deep dive.
Threw the results into a spreadsheet and discovered something interesting that could be worth further investigation. There may be more to crewing than just a bump in average productivity. The standard deviation of the runs for the small ships more than doubled for the crewed runs compared to uncrewed. For those that don't know what that means, standard deviation is a way of measuring the spread of a set of numbers relative to the average. A higher standard deviation means the difference between the highest and lowest numbers is going to be greater. For it to more than double for the small ships means a crewed small ship will have much larger variation between runs than an uncrewed small ship. So crews may only be worth it if you plan to do multiple runs.
On the other hand, the medium ships cut their standard deviation in half by adding a crew. So not only was the average better, their performance was substantially more consistent.
The large ships had their standard deviation increased by crewing as well. But only by about 22%. Would specialists have made a difference in that? I don't know.
Unfortunately, it's a rather small sample set. So there's no telling if those numbers mean anything or if they're just an artifact.
For the trader test you could limit how far they are allowed to jump to one sector and put them in nopileos fortune, that would mitigate a lot of variance while experimenting with trade over distance so that travel time remains a factor.
Thanks! Finally i know that i am not just wasting time with my servicecrew :)
This is superb. Thank you
an idea for handling the variance,
A) large sample size, probably time consuming, but should eat most of the variance
B) fixed or semi-fixed routes, restricting the range to only a few sectors, as I suspect your biggest issue is gonna be gate weirdness, and if crew improve maneuverability, then it's likely they will reduce the weirdness using gates and highways.
C) actually calculating the 95% confidence interval, easy to do in excel, and will tell you if you can exclude no effect, since if the 95% confidence interval includes 0, we don't know if any effect exists
Thank you for answering this question!
Thx for this test!❤
Higher level crew repair faster and to higher hull strength. Better crews also defend the ship better with turrets.
It's now possible to to that test with the new Local trader mechanic just introduced.
Oh yeah! I might have to do a part 2 once the new DLC is released!
Great info..
I kind of just throw service crew everywhere eventually. Most games I'll run two Heracles Vanguards that constantly get fed with rookies and before I know it I have 420 5-star engineers. When it comes to profits-per-hour I could just merc a bunch of Honshu for a quick 30-40 mil per heist if things are moving too slowly for my desired rate of expansion.
i think the best way to test traders, would be to have them on adv autotrade, anchor them to a sector with a shipyard and wharf, and have them only trade ship building components. let them travel a maximum of 1 sector.
this way they wont be flying all over the place. they will be anchored to a small area.
Doesnt this mean that maneuverability is super important even for economy?
Doesnt that mean that investing in the Mk3 combat thrusters will ALSO bring a similar improvement in profit per hour? Even if crewed by amateur crewmen, the increased maneuverability (pitch yaw etc) would increase profits?
This video only focused on the effects of service crew, but what you say is true: better equipped ships will help increase profits
Have you done a t variance test? Id say 40-50% increase must be significant even at such low sample size, but Im not certain.
Hello, thanks for the video. How is the AI worse off with a travel drive vs the all-around version?
You can boost while charging travel mode, which the AI will not do and instead wait out the three seconds. Combined with AI pathing that will fly them into asteroids and reset travel speed it can sometimes be slower in an asteroid heavy sector. The AI has an easier time escaping from Kha'ak and Xenon with all-around drives, which didn't come into play in this video, but is also something to consider
@JK Ninja new player here would love to see this test on all around vs travel if you have one ive been using travel because i figured AI would be slowest going back and forth as you say because they randomly fly across the galaxy sometimes for the asteroid lol
Sweet! this is very helpful, thank you!
Now... I'll feel even worse with a ship full of dozens of people explodes because Kha'ak are a thing.
ALSO. side note, not entirely related to the video. I like how you pronounce Kha'ak and while I get a school boyish chuckle from saying it like "cock" all the time - I think pronouncing it more like Kayak (not exactly how you pronounced it, but close) - is funnier to me.
I had to figure out how to pronounce it, as I'm pretty the way you say it is more correct, but I don't think TH-cam you like that very much 🤣
Just let it fly to a rock and then fire crew in one save to see how much it affects actual mining
I always tend to use 80% crew 20% marines for larger more expensive ships and 100% crew for smaller less expensive.
I did notice on M sized ships with not much crew that inside of them crew members would jump between stations as things happened. Sometimes it didnt effect anything like when traveling and docking, but in combat things seemed more hectic. Of course thats the AI in general when you're in system during combat.
The crew animations inside of a ship are most likely not linked to any performance.
L Miners are much more expensive...and much more durable. More time to react to any Khaak attack. Is that worth paying over 5x as much?
I personally believe so, but I am biased towards large transports
The key difference is in equipment mod compare 10M to 1L for similar % improvememt and u use almost same mods for both (bar turrets). T3 mods on your L miners beats out M miners in overall cost effectiveness IMO. U want drones on Hull.
I've been wondering for a long time if the service crew helps in operations. Any way to test that in combat?
service crew will repair damaged hulls, I believe they will only repair it up to 15-75% depending on engineer skill. On large ships they will also repair damaged modules such as turrets and engines. Is there a particular combat scenario(s) you are interested in testing? If I can get enough of those to make a video I'd be happy to make one!
@@JKNinja Nice thank you.
I've only been playing a few months and noticed that turrets are apparently automtomus. I've flow ships which I accidently didn't have any crew meaning service or marines and had my turrets set on defend. When doing escort mission and came under attack the turrets started defending. Thus I'd assume that service are not as important to combat and marines are only valuable if and when boarding my ship or when I send them to board another.
0:28: "The AI aren't the best with travel engines..."
@JK Ninja Can you clarify that?
The AI will come to a full stop before reengaging the travel drives instead of using fight assist off to maintain momentum. If they need to make a hard turn or flying through asteroids it will severely slow them down
@@JKNinja This only in-sector, or out-of sector too?
@@conradolacerda both, although asteroids aren't rendered in OOS so it isn't as bad there
.........1:19 i see default response to pirates....... ughh....... 100 hours playing i never seen that. they have been the bain of my existance. i loose so many ships to them cause i miss the msg and so my cargo crew says well boss said nothing so i guess we atk them them......... or i have to chase them down to save them time after time......... i wish i seen that 90 hours ago lol
What about combat and repair performance? Do they shoot more accurate with turrets and do they repair engines/modules faster?
The turrets won't shoot faster. Higher level service crew will repair the hull to a higher percentage while repair drones are better for damaged modules.
What if you are OOS?
These tests were done in OOS, the footage was taken afterwards