Development Cycle Archive
Thread: Crafting Critical Failure Rates
BenRod79
Wed Jan 28, 2004 9:12 pm
#235
I am an artisan, as well, and I feel that no matter how hard you work for your rare loot, you should be prepared for the worst. I may only be a Devil's Advocate here, but in the real world, if you were making a metal alloy (for jewelry, for instance) and you misbalanced your metals, you would ruin the whole batch and you would have to scrap it and start over. I have seen jewelers have to melt down and send off thousands of dollars in gold to be processed because of a simple miscalculation on their alloy mixes. I have also seen them lose a gemstone because they got the torch a little too close and cracked the stone.
I do believe that you should (perhaps) be able to salvage some of the resources you use if you have a critical failure, but asking not to lose a resource simply because it is rare is just whining.
That being said, a master goldsmith always double-checks or triple-checks his numbers. His failure rate will not be 4%.
Masters don't make many mistakes.
I do believe that you should (perhaps) be able to salvage some of the resources you use if you have a critical failure, but asking not to lose a resource simply because it is rare is just whining.
That being said, a master goldsmith always double-checks or triple-checks his numbers. His failure rate will not be 4%.
Masters don't make many mistakes.
Masen
Wed Jan 28, 2004 9:26 pm
#236
ok, I have an idea here! YAY! and I actually think it MIGHT be a good one!
Ok..LEAVE critical fails as is, but make a few changes to how the results come out-
On a critical fail...
Any looted items would NEVER vanish. Losing a 500K credit krayt pearl is TOO much of a hit for a Jedi to take. Especially since they cannot place harvesters, etc, so they have no source of fast, huge income to replace them. Similarly..ONE krayt scale can sell for a half mil or more sometimes, that's WAY too much to lose over some arbitrary, and totally unfair system. And no, this has never happened to me, I've never made, or bought anything with that kind of stuff in it, I simply don't have the funds. But, I can definately empathise with people who have.
Crafted component items would have a 50/50 chance of vanishing at Novice level, gradually declining until you never lose crafted components as a master.
Raw materials you would lose all at novice, gradually declining, until you lose half as a master.
The whole idea of this system, is that a master will be making items that take insane amounts of materials, and tons of crafted components. BUT..they would know how to salvage any serious mishap they did make, recoverign their losses as best as they can, though it would still always cost them something, it would never be as devestating. It would also mean that artisan vehicles would not critical..I happen to NOT be in a guild myself, and critical failing ONE swoop bike costs me an entire's day resources just about! That's a BIT too harsh a penalty for a totally random and arbitrary system. Half I could replace..even with a critical, I could make and sell ONE swoop a day. Considering a master marksman can run out, and gun down easy missions for over a hundred K a day if he wanted to..Expecting a master artisan to at least be guaranteed to be able to make about 30-50K a day isn't asking too much.
Naufragus
Wed Jan 28, 2004 9:33 pm
#237
kittle31 wrote:
THAT is what people are griping and complaining about, and that is what needs fixing.
if i had to gripe, what i would gripe about is that i can be using the best materials and best tools. First experiment is GREAT!!!! or even AMAZING! and the product can be at 90%.+..when you go for the second attempt its a CRITICAL FAIL! and it suddenly it is ruined. A critical fail on experimentation SHOULD NOT wreck an item. It should just eat your exeriment points but leave the item at what ever level it was at
basically as you said there is a lot of stuff that just has to be thrown out. I throw anything away thats just gets a GOOD!.
Cybst0rm
Wed Jan 28, 2004 10:14 pm
#238
I didn't read all of the posts, so if someone already said this, please ignore my ramblings....
It's a scary thought when people thing that a 4% failure rate isn't that bad. Let me break it down for ya'll.
4% crit rate means 1 in 25 times you experiment, YOU WILL GET A CRITICAL FAIL. So lets say I am experimenting on some Composite Armor chest pieces. I plan to make 5 of them. On the combines, odds are pretty good that I won't get a critical fail. Now.... lets talk about experimentation. Each time you click "Experiment" that counts towardsme 1 in 25. Statistically, I WILL crit fail on at least 1 Composite Chest piece while experimenting. Guess what happenes to that chest piece? You guessed it... I hit the "X" at the top and start over.
So here is my idea. Reduce the amount of experimentation points but increase the effect they have on the finished product. Instead of having 10 points, typically, you get 5 but 1 point will do twice the effect as the other way. That means your chance to fail will remain the same, yet the frequency will decrease because you are able to push out more products before the evil crit monster rolls around. Granted, that will decrease some flexibility in crafting, although I don't think that is such a big deal since each crafter has their target goal. As an Armorsmith, I max out quality first to increase base resists on armor. Next, I go for durability to decrease the HAM cost. Maxing out quality typically take me 7 experimentation points and 3 tries. Giving us less points means I can attain that same quality % with 3-4 points in only 1-2 tries. That still leaves me with a decent amount of points to put into other things.
Also, crit fails need to be scaled a bit. Before I do a combine or experiment, I always chuckle when it says 0% risk. Why? I know it doesn't mean anything. Why did I bust my butt to collect resources to make the finest tools (+14.6) and crafting stations (+41) to make it mean nothing? It's time to bring those stats into focus a little more. Here is the problem though... SOE needs to make having quality tools mean something. Why you ask? Becuase in a real economy, not everyone can afford the Ferrari. In SOE's economy, any n00b with 30 minutes of free time can. So how about putting in a graduated scale of bonuses based on the quality of tool/station used? A Master knows how to use his tools much better then a Novice just starting out. This would reflect that and give some realism to gaining skills. Even better... give a + to experimentation! Each elite profession has 17 skill boxes. For top quality weapons, you could add a certain percentage to getting another or "extra" experimentation point. Base or poor quality tools and stations would not confer any or even take away from experimentation points. Granted, this is a pretty dynamic solution so we may never see something like this. I feel it might be to much for any MMORPG to pull off, computer resource wise. Just a thought though.
Vorg
Wed Jan 28, 2004 10:29 pm
#239
After all teh work we put into getting good resources with this crapy resource system, getting good quality tools, and working up the skill tree, our work should be distroyed by these falures, ether in assembly or this expermentation thing. The resualts of our work should be based on the resources, tools and our skill. not some stinking random generator. it ruins the game.
Quasar64
Thu Jan 29, 2004 12:02 am
#240
I completely agree with Cybst0rm ... these 4+%s can really add up.
Assembly attempts: 1448
Assembly critical failures: 65
Assembly critical failure rate: 4.49%
Experimentation attempts: 1284
Experimentation critical failures: 54
Experimentation critical failure rate: 4.21%
4+% chance to crit fail the assembly of an item.
4+% chance to fail on EACH experimentation point?
I am also a Master Armorsmith. I dont recall how many times I experiment on each piece of armor, but it's quite a bit. I think most players experiment with a single "point" at a time in any given experimentation row.
Is 4.21% the chance to fail on EACHpoint in an experimentation row?
or
Is there one "Experimentation Attempt" per crafted item?
Thanks
Awis
Thu Jan 29, 2004 1:30 am
#241
Ok i didnt read through all 10 pages so this may have been posted, but how can i fail while experimenting when there is 0% risk?
LadyGrey
Thu Jan 29, 2004 1:32 am
#242
I was just thinking about critical failures in relation to freeway overpasses. I grew up in S. California, and freeway overpasses are all over the place, so they came to mind. If a 4% critical failure rate is acceptable, then that would mean that, regardless of the skill of the engineers and builders, and quality of materials, 1 in 25 of those freeway overpasses should just simply collapse right after being put up. Or, put another way, 1 in 25 buildings should just collapse right after being built.
The point is, that skill and good materials makes a big difference on whether something can be considered reliable. There are always going to be unexpected problems, but skill and good materials can even reduce those, to some extent. I can see the point of having critical failures when you are building up skills, but then again, you aren't building things that are really big and complicated, either. I can even see having some very small amount of critical failures while experimenting, because experimenting implies that you are taking some chances. However, having skill and good materials should even make those a very rare occurrence. Even when I am experimenting with something, my skill gives me the ability to control that experimentation to maximize the positive aspects of it.
I don't see why I should have to experiment on 30 BEC's to get one really good one, using good resources, a good tool, and a good crafting station. I would suggest that a master-class artisan be allowed a special crafting mode, one in which they can focus intently on the job at hand, perhaps taking three or four times longer to do the crafting, but with a much higher chance of getting an amazing result from the experimentation.
The point is, that skill and good materials makes a big difference on whether something can be considered reliable. There are always going to be unexpected problems, but skill and good materials can even reduce those, to some extent. I can see the point of having critical failures when you are building up skills, but then again, you aren't building things that are really big and complicated, either. I can even see having some very small amount of critical failures while experimenting, because experimenting implies that you are taking some chances. However, having skill and good materials should even make those a very rare occurrence. Even when I am experimenting with something, my skill gives me the ability to control that experimentation to maximize the positive aspects of it.
I don't see why I should have to experiment on 30 BEC's to get one really good one, using good resources, a good tool, and a good crafting station. I would suggest that a master-class artisan be allowed a special crafting mode, one in which they can focus intently on the job at hand, perhaps taking three or four times longer to do the crafting, but with a much higher chance of getting an amazing result from the experimentation.
Gron_DM
Thu Jan 29, 2004 1:38 am
#243
i get more then 4%, i have made around 50 or so 4BER harvesters, (or less then 4BER...) and out of them i have had 10 3ber that is 1/5th or 20% that had 1 or more crit fails....tools: 14.2, craftstation: 42.93, master artisan. materials averaged a high enough qual (SR,UT and HR) to acheive a 80% or a 4BER. this is par for the course in MDE also i make BH droids and after making over 50 runs of parts i have seen 10 crit fails in experimentation there too. materials and sup-comp quals vary from 40% to 70% for the GP and mem modules but point is i am a master DE.
SwyftHaE
Thu Jan 29, 2004 1:41 am
#244
Master Doc/ Master CM, and I still fail all the time, it's nuts. Even if you use the best tools, the best resources, you have a good chance to fail. I have to experiment several times to finish a product sometimes, so I have about a 20-30% chance to fail everytime I craft. There's something wrong there when you are a MASTER and you have all the best items you can muster up and you still fail. I would understand if you were a novice medic and crafted up something just to grind points and you failed, but it's a bit tedious if you're a master failing all the time.
Revamp the crafting tree!
atimes
Thu Jan 29, 2004 1:45 am
#245
Assembly attempts: 1448
Assembly critical failures: 65
Assembly critical failure rate: 4.49%
Experimentation attempts: 1284
Experimentation critical failures: 54
Experimentation critical failure rate: 4.21%
Assembly critical failures: 65
Assembly critical failure rate: 4.49%
Experimentation attempts: 1284
Experimentation critical failures: 54
Experimentation critical failure rate: 4.21%
No offensedevs but your metrics and your numbers don't give you a proper sense of the problem. Until you experience or get a proper sense of the sheer frustruation of a critical failure on a final combine losing thousands of resources and components that took you a full solid real world week to collect you will not be able to understand just how devasting that 4% failure rate is. It's not the small percentage of failure. It's the failure causing you to lose thousands of resources that you will never be able to get back.
Ipersonally know of architects who have experienced immense levels of real life human stress because of crit failures on PA Halls and other large structures. If I get a crit failure on a basic camp kit it's no big deal. If I get a crit failure on a PA hall I'm likely to go nuts.
I can not tell you the amount of posts by architects and other crafters who basically said "critical failures are too stressful and this is a game. Whenver this game feels like work it's time to try something else. I'm quitting <insert crafting profession here> and picking up a gun to try to have some fun"
That issue IMO is what needs to be addressed. Critical failures while realistic (and adding to overall game feel) well. . .stink and for some theystink really bad.
Lothario_Gray
Thu Jan 29, 2004 1:58 am
#246
I am going to be hated for this post. I have read throughout about a handfull of these posts and come up with the same conclusions.
People are grinding to master and than expect never to fail once they get there. They have no problem failing if it is during grinding, but will not tolerate losing a 20 million credit item later on. Some professions should get better than others at crafting, or worse. People are also whining because if they do not get all master successes, they can't use the item. Oh and everyone is whining because it is not perfect for them. People need all levels of weaponry. From the single skill boosting power up to a weak CDEF pistol. There is a market, you just need to price the items correctly.
Ok, Now you can hate me. Early on in developement, there was mention to having an item be easier to make at the level that you get the schematic at. I do not think this is the case anymore. I think the failure rate should rise and fall based on how close you are to the schematic in question Hence a novice Artisan will be able to make some pretty good CDEF Pistols, but a Master artisan will not.
I think that critical failures are just fine the way they are. So you loose some items. Does it really mater, you will get them back within an hour or two anyways.
Hawkrw
Thu Jan 29, 2004 2:23 am
#247
I pretty much figured it was a 5% crit failure rate because of the practice mode returns, so your numbers don't suprise me. And in all honesty I probably see about 5% failure rate. The problem for me as a master armorsmith and droid engineer is as follows.
- 5% failure applies to crafting
- 5% failure applies to each experimentation attempt
- On avg suit of armor I do at least 4 seperate experiments. This is a 19 in 20 chance at least 5 times every time I craft a single piece. Then you take into account I have to craft 9 pieces for most suits of armor and I'm rolling the dice 45 times on a 19 in 20 chance. Statistically I'm gonna see at least 2, if not 3 failures on avg (take it to infinity and the expected failure rate would be 2.5 failures per suit of armor). Now a single failure ruins the piece of armor. So I'm lossing 3 pieces every time I make armor. Mind you now I'm looking at a 1/3 loss rate, not 5% loss rate. Take into account the loss rate on layers and segments as well and it's probably closer to 1 in 4 loss rate of my resources. So I can basically kiss goodbye to 1/4 of my resources in the long run. That's a bit extreme, particularly when you're working with looted items. I am probably dreaming to get the 1 in 20 on the entire project but even if it's fixed so I check once on each comp and once on the final product it'll drop it down to something more along lines of a 10% resource loss, not 25%.
-Now for droids the impact is a lot less. Here you can usually only experiment a module up .5-1 notch and even then it takes best resources. Mech quality doesn't seem to do much but I'll admit I dable in droids more then mass produce. So for most part you either experiment or you don't, if it fails you just have the module you started with. If you don't want to deal with the risk just make a mod 6 and you have a single check (assembly). This is likely to change though when droids are revamped, if it's anything like bio-engineer this will be an issue much like the armor one Right now biggest problem with DE is that you can lose a lot of components if you fail at the end.
I imagine the other crafting professions fall into one of the above 2 templates. Where you either risk it all about 1 in 3 or 1 in 4 times or you just risk a lot of components and work at one time.
Ideas for improvement
One) Move experiment checks to the end result rather then every experiment. In this instance a great gets exactly what you wanted, an amazing is even better (10% boost), a good is 10% less, a mod is 20% less, a basic success is 30% less, so forth to a crit failure being a 50% less result. In this case an amazing is worth something (isn't funny how you always get'em on that last box and it does nothing) and a crit failure will be a predictable outcome.
Two) Retreval, I'm mixed on this one. For one the biggest risk is during experimenting so we'd need a "scrap project" option. Second, I don't know how much the game tracks the resources and components used but I imagine maintaining them beyond the early stages may take some re-working of the system.
Three) Flat check on final product. This would go after the create prototype, experiment, schematic option. Basically if you choose create prototype or schematic the game checks to see if it crit failed, if you didn't congrats you've a new item, if it did you're starting over. This would guarntee 5% per every item no matter what you did to it. You could also keep the assembly stage and experiment stage the same just remove the crit failure part, regular failures and varrying success have same results as before.
Anyways, quick recap. The problem isn't the actually % chance of a failure, rahter it's the frequency in which the % is checked. Statisitcally you'll end up failing in a master type product about 1 in 4 times just because you're checking it that many times. Since near perfection is required in most master products this renders the item useless and scrap. To fix this we either need a way to try again without lossing everything or to have the check fixed to the proper 1 in 20 it's intended to be.
- 5% failure applies to crafting
- 5% failure applies to each experimentation attempt
- On avg suit of armor I do at least 4 seperate experiments. This is a 19 in 20 chance at least 5 times every time I craft a single piece. Then you take into account I have to craft 9 pieces for most suits of armor and I'm rolling the dice 45 times on a 19 in 20 chance. Statistically I'm gonna see at least 2, if not 3 failures on avg (take it to infinity and the expected failure rate would be 2.5 failures per suit of armor). Now a single failure ruins the piece of armor. So I'm lossing 3 pieces every time I make armor. Mind you now I'm looking at a 1/3 loss rate, not 5% loss rate. Take into account the loss rate on layers and segments as well and it's probably closer to 1 in 4 loss rate of my resources. So I can basically kiss goodbye to 1/4 of my resources in the long run. That's a bit extreme, particularly when you're working with looted items. I am probably dreaming to get the 1 in 20 on the entire project but even if it's fixed so I check once on each comp and once on the final product it'll drop it down to something more along lines of a 10% resource loss, not 25%.
-Now for droids the impact is a lot less. Here you can usually only experiment a module up .5-1 notch and even then it takes best resources. Mech quality doesn't seem to do much but I'll admit I dable in droids more then mass produce. So for most part you either experiment or you don't, if it fails you just have the module you started with. If you don't want to deal with the risk just make a mod 6 and you have a single check (assembly). This is likely to change though when droids are revamped, if it's anything like bio-engineer this will be an issue much like the armor one Right now biggest problem with DE is that you can lose a lot of components if you fail at the end.
I imagine the other crafting professions fall into one of the above 2 templates. Where you either risk it all about 1 in 3 or 1 in 4 times or you just risk a lot of components and work at one time.
Ideas for improvement
One) Move experiment checks to the end result rather then every experiment. In this instance a great gets exactly what you wanted, an amazing is even better (10% boost), a good is 10% less, a mod is 20% less, a basic success is 30% less, so forth to a crit failure being a 50% less result. In this case an amazing is worth something (isn't funny how you always get'em on that last box and it does nothing) and a crit failure will be a predictable outcome.
Two) Retreval, I'm mixed on this one. For one the biggest risk is during experimenting so we'd need a "scrap project" option. Second, I don't know how much the game tracks the resources and components used but I imagine maintaining them beyond the early stages may take some re-working of the system.
Three) Flat check on final product. This would go after the create prototype, experiment, schematic option. Basically if you choose create prototype or schematic the game checks to see if it crit failed, if you didn't congrats you've a new item, if it did you're starting over. This would guarntee 5% per every item no matter what you did to it. You could also keep the assembly stage and experiment stage the same just remove the crit failure part, regular failures and varrying success have same results as before.
Anyways, quick recap. The problem isn't the actually % chance of a failure, rahter it's the frequency in which the % is checked. Statisitcally you'll end up failing in a master type product about 1 in 4 times just because you're checking it that many times. Since near perfection is required in most master products this renders the item useless and scrap. To fix this we either need a way to try again without lossing everything or to have the check fixed to the proper 1 in 20 it's intended to be.