To add another comment about why this is frustrating/upsetting: since the distribution of results widens the further you get from 99% success AND the log based formula for success reduced the incremental value of each crew, then there is nearly no point to buy packs, spend time optimizing shuttles and fighting for incremental value of you are talking about the difference between 75% and an 82% chance of winning.
In this case, nstead of crew, skill and strategy defining who places top 1k, then RNG becomes the dominant factor. This makes a huge difference in how much effort and time I put into the shuttles if my planning doesn’t really matter
My theory is as follows. I’ve staffed shuttles doing the average skill calcs myself in excel and using crew with first stat primary on and slots and I’ve had results in line to slightly above the displayed percentages the last few events.
Observation:
The displayed percentage caluculates average skill for AND missions as HIGH skill plus 1/4 LOW skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci and a character with 1000 sci and 400 dip will both have a 1100 average skill).
Theory:
The actual (server side percentage) calculates as FIRST skill plus 1/4 SECOND skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci would have an 1100 average skill but a 400 dip 1000 sci character would have a 650 average skill)
This is exactly what I have been doing for the past few months and it has been working out for me very well. My only difference is that I was using formula FIRST skill + 1/2 Second skill. Any mission with stats above 5000 will pass most of the time and closer to or over 7000 have about 98% success rate despite display rate. I run the same missions with the same cards in the same slots once I hit 3 hr. If I fail one of the four, it's usually the lowest one. I do try to put cards main skill in first then lower skill after AND. I stay away from OR slots.
My theory is as follows. I’ve staffed shuttles doing the average skill calcs myself in excel and using crew with first stat primary on and slots and I’ve had results in line to slightly above the displayed percentages the last few events.
Observation:
The displayed percentage caluculates average skill for AND missions as HIGH skill plus 1/4 LOW skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci and a character with 1000 sci and 400 dip will both have a 1100 average skill).
Theory:
The actual (server side percentage) calculates as FIRST skill plus 1/4 SECOND skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci would have an 1100 average skill but a 400 dip 1000 sci character would have a 650 average skill)
I wish this was the issue for me, but last event I only had one crew slot that wasn’t already highest first, lowest second. I will try to optimize it that way this event, but there is potentially a huge opportunity cost if you are wrong. Again, it would be awesome to know how the server is actually calculating this.
To add another comment about why this is frustrating/upsetting: since the distribution of results widens the further you get from 99% success AND the log based formula for success reduced the incremental value of each crew, then there is nearly no point to buy packs, spend time optimizing shuttles and fighting for incremental value of you are talking about the difference between 75% and an 82% chance of winning.
@Shan , @Black Pebble please take this feedback to the powers that be. We as players want to keep playing this game, and some still even want to spend money.
But if these packs are bugged or ineffective, as GTMET says, there is no point to buying these packs.
Lack of response/addressing this issue will result in a loss in revenue for DB. Please dont let it come to that.
Oh, and just to update everyone, CS did respond to my ticket claiming that this may be due to a client/server lag issue and that my phone being old/out of memory/slow/etc. They suggested a reboot of the game.
Obviously, that didn't work.
Mind you, I'm on a Samsung Galaxy S9+, which by definition is new, fast, and has the most memory (6gb).
Assuming the reports here about CS responses came from well-informed staff and are therefore not bogus, one starts to wonder - what's the frigging point of messing with the global success/failure rate at all? If EVERYONE is affected, that rate shouldn't matter at all, and will only aggravate people, as they apparently and consistently artificially force a much lower rate than any of the rates displayed.
The only possible reason that springs to mind is that they are trying to force people to invest more by artificially putting the breaks on the average returned VP (at least during events), trying to crank up the competition. Which would expose DB yet again as just a revenue-obsessed company caring about nothing else (so no surprise there :-/ ).
However, even this argument is extremely weak, if not nonsensical: if EVERYONE is equally effected (i.e., the rate is lowered by the same amount for everybody), adjusting this will only have an effect on the outcome for the ABSOLUTE amount of VPs, not the relative amount. However, the only place where the absolute amount has any effect are the threshold rewards - and the total amount of VP required here is so low that most people achieving the last 4* reward here (=usually the top reward) will not even have to invest any DIL (let alone money) anyway.
For all ranked rewards, nothing will change, even is the success rate is artificially lowered by half (or even doubled, for that matter) (assuming every player is continuously playing throughout the event).
I doubt there is an actual global effect. I suspect the CS reps didn't understand the question, don't speak engrish as a first language, crafted a very poorly written response, and just respam out the canned form letter.
I know for a fact that a great many tickets are not read. At best they're glanced at and given a summary judgement based off of 1-3 key words they find. I and my entire fleet have had CS reps respond to tickets asking for information provided within the first 8 words of our tickets, asked for information already provided in screen shots, and required multiple responses to get them to read anything.
However, if there is an actual global slow down, that's the epitome of idiocy. All you're going to do is anger players, drive them away out of frustration, and create these threads over and over again.
Perfectly within DB's history, so maybe . . .
@shan it would be great if someone from DB would chime in on this speculation and interpretation of CS feedback:
1) In events, are each players shuttles based only on their crew and difficulty, or is there some global effect that reduces the actual success rate?
2) if the displayed success rates are intended to be accurate, would you someone from DB commit to working with the small number of players who have all recorded data that shows that event shuttles are nearly certainly not achieving the displayed success rate
A reply and good faith effort would go a long way in restoring faith in the fairness and transparency of the event system.
....are you turning blue yet?....
Founding ADM - PoF family of fleets (POF, POF2 & POF3) - Dear TP: Non sequitur. Your facts are uncoordinated.
I always thought the percents didn't matter if they use a RNG to decide it.
Trista
A typical way of doing it in code would be to do rand(100) - giving a random number between 0 and 100. If the random number is less than the percentage likelihood it’s a success, else fail.
IIRC, that includes zero and 100, which gives you 101 numbers that might be chosen. Which would explain how 0%ers succeed, if it's more of a if randNumber > successRate, shuttle = fail, rather than if successRate < randNumber, shuttle = succeed. So that in the case that randNumber = successRate, it succeeds. So 0 percenters succeed a little under 1% of the time.
I do wonder if the server's method of calculating percentages differs from the way the UI calculates and displays them, though. It could make sense if they use a more complicated equation on the server side than on the UI side, in order to allow for faster calculations of percentages.
Just a thought.
Please note: It's been years since I even thought about doing more coding than an Excel/Google Sheets workbook, or a Wiki edit. I'd like to request you go gentle on me in the case that I am completely and irrevocably wrong. (Which is all too likely. )
I think you're right with the typical unix/c "rand()", but it does depend on the language etc. Some of them return integers, some return floats. If its a float, you end up with somewhere between 0 and 100 inclusive, but you wont get 100.1 for example. then you end up piping it through an int() or roundup/down function.
However the kind of problematic discrepancies people are reporting arent really the odd +/- 1%, but sometimes way off. I think there probably is some difference between the maths shown to the user vs the logic on the server as you describe. If what @Nomad (TP:TG) described is true that would certainly explain it, but then its a bit unfair to describe it as a "% success".
I just ran my first "normal" 3hr shuttles with 90min reduction boosties, and they ALL showed clearly 99% ... guess what, I had two ... TWO failz ... that's 50% and not 99% at all. Instead of sending 2250 VP 9hr-overnights, I am down to 1500 VP now. Which means Top 1000 is already very unlikely, even with bonus points. I will be too far behind, so this event is officially over for me.
"Everything about the Jem'Hadar is lethal!" - Eris (ST-DS9 Episode 2x26 "The Jem'Hadar")
My theory is as follows. I’ve staffed shuttles doing the average skill calcs myself in excel and using crew with first stat primary on and slots and I’ve had results in line to slightly above the displayed percentages the last few events.
Observation:
The displayed percentage caluculates average skill for AND missions as HIGH skill plus 1/4 LOW skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci and a character with 1000 sci and 400 dip will both have a 1100 average skill).
Theory:
The actual (server side percentage) calculates as FIRST skill plus 1/4 SECOND skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci would have an 1100 average skill but a 400 dip 1000 sci character would have a 650 average skill)
@Paladin 27, if you have a ENG AND DIP slot and you put a crew in with only 1000 DIP. How do you think the game manages it, as a 1000 or as 1000/4 = 250?
My theory is as follows. I’ve staffed shuttles doing the average skill calcs myself in excel and using crew with first stat primary on and slots and I’ve had results in line to slightly above the displayed percentages the last few events.
Observation:
The displayed percentage caluculates average skill for AND missions as HIGH skill plus 1/4 LOW skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci and a character with 1000 sci and 400 dip will both have a 1100 average skill).
Theory:
The actual (server side percentage) calculates as FIRST skill plus 1/4 SECOND skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci would have an 1100 average skill but a 400 dip 1000 sci character would have a 650 average skill)
@Paladin 27, if you have a ENG AND DIP slot and you put a crew in with only 1000 DIP. How do you think the game manages it, as a 1000 or as 1000/4 = 250?
Going to track my shuttles both ways this event
My theory is that on screen it treats it as 1000, but server side 250.
My theory is as follows. I’ve staffed shuttles doing the average skill calcs myself in excel and using crew with first stat primary on and slots and I’ve had results in line to slightly above the displayed percentages the last few events.
Observation:
The displayed percentage caluculates average skill for AND missions as HIGH skill plus 1/4 LOW skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci and a character with 1000 sci and 400 dip will both have a 1100 average skill).
Theory:
The actual (server side percentage) calculates as FIRST skill plus 1/4 SECOND skill (e.g. for a dip and sci slot a character with 1000 dip and 400 sci would have an 1100 average skill but a 400 dip 1000 sci character would have a 650 average skill)
@Paladin 27, if you have a ENG AND DIP slot and you put a crew in with only 1000 DIP. How do you think the game manages it, as a 1000 or as 1000/4 = 250?
Going to track my shuttles both ways this event
My theory is that on screen it treats it as 1000, but server side 250.
Well I should have a good test case this event, a couple of my shuttles look great either way and a couple are terrible if the server side meets your criteria....
Possible the 2nd is always the 25% theory that Paladin suggested is correct. Engineering is the first trait for all these slots. That would mean that your traits are being calculated as
235
161
281
not
942
644
1123
Your success rate would be closer to 60% with those stats, even with the event crew bonus.
I’m assuming it’s that three-slot all ENG missions (but with no & SCI) “Adjusting the Emitters” because that’s what it looks like... and you didn’t send a single ENG character and went with strong characters in the secondary skills of CMD/DIP/SEC... maybe that has something to do with it? Dunno... but a 300vp mission should be like an instant win even with crippled 1* crew...
I had a great kickstart, clearing 25 missions in about 10 minutes. I then went 4/4 on the 1800 point missions, with time boosts... everything was going well!
However, I am now repeating the 2250 point wave of missions for the third time! On the first 2 waves I went 3/4 and 2/4, despite not having a mission rank below 94% (all with time boosts).
Note: I rant the "Emitters" mission during each wave, without covering any of the ENG skills, as I went with the crew giving the best displayed result. Part of me wonders if that particular mission is not displaying accurately (though the second skill was extremely high in all three seats), but I don't think I failed that one twice, if at all.
based on the feedback I am seeing, I am just waiting for that to start failing on me as I only have one engineer in any of the three slots (even if it is a FF/FE romulan data)
based on the feedback I am seeing, I am just waiting for that to start failing on me as I only have one engineer in any of the three slots (even if it is a FF/FE romulan data)
Im running Rom Data 2/5, T'Rul and Borg Queen on it... waiting to see when my ENG luck runs out.... But BQ has been carrying slot three for me into 2750 so we shall see.....
As I've mentioned in other threads, I started collecting data on shuttle displayed percentages vs. results, at the single percent interval. I've been collecting data for just over 2 months now, and have over 1400 data points.
So far, the overall average is 1.1% below what's displayed. However, the data is still far from sufficient to make any real conclusions (I consider 100 data points on a single percent displayed to be a sufficient sample size, for which I only have at one value).
I'll put the link here again, for anyone who would like to help contribute: https://goo.gl/forms/3uLNW2tuLYikAMSE3
If you've been collecting your own data, I'm willing to give access to the back-end for larger data dumps. However, please note that I'm tracking by individual percentages; aggregate results are not useful.
As I've mentioned in other threads, I started collecting data on shuttle displayed percentages vs. results, at the single percent interval. I've been collecting data for just over 2 months now, and have over 1400 data points.
So far, the overall average is 1.1% below what's displayed. However, the data is still far from sufficient to make any real conclusions (I consider 100 data points on a single percent displayed to be a sufficient sample size, for which I only have at one value).
I'll put the link here again, for anyone who would like to help contribute: https://goo.gl/forms/3uLNW2tuLYikAMSE3
If you've been collecting your own data, I'm willing to give access to the back-end for larger data dumps. However, please note that I'm tracking by individual percentages; aggregate results are not useful.
I think we might need to take a look at what you (and for that matter what I) am/are doing differently (if anything) than folks who are seeing wide variance. I'm in the same boat you are, not seeing a ton of fails, just occasional ones, and in the range of what would be expected.
As I've mentioned in other threads, I started collecting data on shuttle displayed percentages vs. results, at the single percent interval. I've been collecting data for just over 2 months now, and have over 1400 data points.
So far, the overall average is 1.1% below what's displayed. However, the data is still far from sufficient to make any real conclusions (I consider 100 data points on a single percent displayed to be a sufficient sample size, for which I only have at one value).
I'll put the link here again, for anyone who would like to help contribute: https://goo.gl/forms/3uLNW2tuLYikAMSE3
If you've been collecting your own data, I'm willing to give access to the back-end for larger data dumps. However, please note that I'm tracking by individual percentages; aggregate results are not useful.
I also some someone else track data and show similar 10% gaps over time.
Likewise, If you were able to add my data from the last event, your data would get pummeled and probably look more in line with everyone else's data
My wife has been losing 1800 difficulty shuttles 2/4 and 3/4 with All FE event and bonus crew. She has been ignoring the feedback around AND nodes and putting all here skilled crew in the second half of the AND node with engineering in the front half, and boosting engineering. The game is displaying 99% shuttles and she has lost close to 1/3rd
Put another way, a shuttle with all three spots boosted running FF Mirror Spock, Assimilated Tuvok and Romulan Kirk should not lose back to back 1800/2250 shuttles.
I believe there is a systematic bug occuring (although it might not be this AND node issue). Because it seems that certain crew/shuttle combinations will work for one event, all coming in around the correct %, and then the next event everything bombing to pieces. I also have never had an event where I miraculously won 99% of my shuttles, which should happen as well.
I believe the AND node piece is at least part of the issue. Otherwise I would not be in the good shape I am in, as I cover all the skills in an AND even if that drops my displayed percentage. There may be other issues, but if that is truly a factor, it ain't helping things.
Comments
In this case, nstead of crew, skill and strategy defining who places top 1k, then RNG becomes the dominant factor. This makes a huge difference in how much effort and time I put into the shuttles if my planning doesn’t really matter
This is exactly what I have been doing for the past few months and it has been working out for me very well. My only difference is that I was using formula FIRST skill + 1/2 Second skill. Any mission with stats above 5000 will pass most of the time and closer to or over 7000 have about 98% success rate despite display rate. I run the same missions with the same cards in the same slots once I hit 3 hr. If I fail one of the four, it's usually the lowest one. I do try to put cards main skill in first then lower skill after AND. I stay away from OR slots.
I wish this was the issue for me, but last event I only had one crew slot that wasn’t already highest first, lowest second. I will try to optimize it that way this event, but there is potentially a huge opportunity cost if you are wrong. Again, it would be awesome to know how the server is actually calculating this.
But if these packs are bugged or ineffective, as GTMET says, there is no point to buying these packs.
Lack of response/addressing this issue will result in a loss in revenue for DB. Please dont let it come to that.
Obviously, that didn't work.
Mind you, I'm on a Samsung Galaxy S9+, which by definition is new, fast, and has the most memory (6gb).
I doubt there is an actual global effect. I suspect the CS reps didn't understand the question, don't speak engrish as a first language, crafted a very poorly written response, and just respam out the canned form letter.
I know for a fact that a great many tickets are not read. At best they're glanced at and given a summary judgement based off of 1-3 key words they find. I and my entire fleet have had CS reps respond to tickets asking for information provided within the first 8 words of our tickets, asked for information already provided in screen shots, and required multiple responses to get them to read anything.
However, if there is an actual global slow down, that's the epitome of idiocy. All you're going to do is anger players, drive them away out of frustration, and create these threads over and over again.
Perfectly within DB's history, so maybe . . .
....are you turning blue yet?....
I think you're right with the typical unix/c "rand()", but it does depend on the language etc. Some of them return integers, some return floats. If its a float, you end up with somewhere between 0 and 100 inclusive, but you wont get 100.1 for example. then you end up piping it through an int() or roundup/down function.
However the kind of problematic discrepancies people are reporting arent really the odd +/- 1%, but sometimes way off. I think there probably is some difference between the maths shown to the user vs the logic on the server as you describe. If what @Nomad (TP:TG) described is true that would certainly explain it, but then its a bit unfair to describe it as a "% success".
@Paladin 27, if you have a ENG AND DIP slot and you put a crew in with only 1000 DIP. How do you think the game manages it, as a 1000 or as 1000/4 = 250?
Going to track my shuttles both ways this event
The technology was too foreign to you and most of your crew. You were unable to assist in repairing the vessels.
Clearly! /s
My theory is that on screen it treats it as 1000, but server side 250.
Well I should have a good test case this event, a couple of my shuttles look great either way and a couple are terrible if the server side meets your criteria....
Oh snap. Peachtree's got jokes.
Possible the 2nd is always the 25% theory that Paladin suggested is correct. Engineering is the first trait for all these slots. That would mean that your traits are being calculated as
235
161
281
not
942
644
1123
Your success rate would be closer to 60% with those stats, even with the event crew bonus.
I’m assuming it’s that three-slot all ENG missions (but with no & SCI) “Adjusting the Emitters” because that’s what it looks like... and you didn’t send a single ENG character and went with strong characters in the secondary skills of CMD/DIP/SEC... maybe that has something to do with it? Dunno... but a 300vp mission should be like an instant win even with crippled 1* crew...
However, I am now repeating the 2250 point wave of missions for the third time! On the first 2 waves I went 3/4 and 2/4, despite not having a mission rank below 94% (all with time boosts).
Displayed: 96%, 96%, 96%, 96%, 95%, 95%, 94%, 94% = 95.25% average displayed
Actual: 5/8 = 62.5% result
Difference: 32.75% lower than displayed!!!
Note: I rant the "Emitters" mission during each wave, without covering any of the ENG skills, as I went with the crew giving the best displayed result. Part of me wonders if that particular mission is not displaying accurately (though the second skill was extremely high in all three seats), but I don't think I failed that one twice, if at all.
Im running Rom Data 2/5, T'Rul and Borg Queen on it... waiting to see when my ENG luck runs out.... But BQ has been carrying slot three for me into 2750 so we shall see.....
So far, the overall average is 1.1% below what's displayed. However, the data is still far from sufficient to make any real conclusions (I consider 100 data points on a single percent displayed to be a sufficient sample size, for which I only have at one value).
I'll put the link here again, for anyone who would like to help contribute: https://goo.gl/forms/3uLNW2tuLYikAMSE3
If you've been collecting your own data, I'm willing to give access to the back-end for larger data dumps. However, please note that I'm tracking by individual percentages; aggregate results are not useful.
I think we might need to take a look at what you (and for that matter what I) am/are doing differently (if anything) than folks who are seeing wide variance. I'm in the same boat you are, not seeing a ton of fails, just occasional ones, and in the range of what would be expected.
Several other people also tracking data came up with significantly lower percentages. FF sisko showed this with around 1000 event shuttles:
https://forum.disruptorbeam.com/stt/discussion/1713/shuttle-mission-success-chances-during-events-post-your-data-here#latest
I also some someone else track data and show similar 10% gaps over time.
Likewise, If you were able to add my data from the last event, your data would get pummeled and probably look more in line with everyone else's data
My wife has been losing 1800 difficulty shuttles 2/4 and 3/4 with All FE event and bonus crew. She has been ignoring the feedback around AND nodes and putting all here skilled crew in the second half of the AND node with engineering in the front half, and boosting engineering. The game is displaying 99% shuttles and she has lost close to 1/3rd
Put another way, a shuttle with all three spots boosted running FF Mirror Spock, Assimilated Tuvok and Romulan Kirk should not lose back to back 1800/2250 shuttles.
I believe there is a systematic bug occuring (although it might not be this AND node issue). Because it seems that certain crew/shuttle combinations will work for one event, all coming in around the correct %, and then the next event everything bombing to pieces. I also have never had an event where I miraculously won 99% of my shuttles, which should happen as well.