Update raw loot % displays

savacc said:

Sometimes ya just gotta be smarter than the tool you use. The true masters, of anything, are the ones who know when the tool is wrong.

Click to expand...

Not a lot of master carpenters keeping a level that doesn't work in their toolbox. When they realize it is inaccurate, they fix or replace it. We should too

When it came out we were told it was effectively an excel file with rough estimates put in for the ranges. A good example here is 8 guns, one of the most popular RE'd part anywhere else in JtL. Decent w8 masses are not being kept because the numbers used for w8 were picked from a higher tier than other parts. You can see this in action using the powerbi app what would normally be a keeper rarity is showing as 90% or so, and 95% is listed as 17xx RED and 23xxx rather than the more reasonable 18xx red and 25xxx. This has had an extreme impact on the RE market for 8 guns, and lead people to largely using the crafted borstels instead because getting into a starter 8 gun is so hard. The inverse is happening on a lot of other parts, where 95% was set far to low, and people think they have something really rare at 96, or 97%, when it's actually a fairly common drop, trade space is full of these, not to mention not having any indication to show if a part is reward or outclassed by crafted. The over 100% issue is caused by armors and armor stats having a huge range and the file not accounting for it, the mass on a 10 armor can reasonably drop anywhere from 50k to -2k. I estimate the actual range was set to something like 20k-10k

Give pilots a week with that excel file and we can get it to a much more usable state. Because it's just a read text file, backwards compatibility won't be a problem.

-Edit- I was not prepared for just how dry the market is for w8s, and I 100% blame the % system scarying people off of them, thinking it's only keepable at extremely rare masses.

Not sure what you are referencing, but the space loot system is just a bell curve. Each named part has a mean and a standard deviation hard coded, eg Quantum Ion Drives have a Top Speed mean of 92, with a standard deviation of 4.6. This means that a QID with a top speed of 101.2 is 2 standard deviations away from the mean, with a 2.5% chance of occurring. Going out to 3 standard deviations puts you at 105.8, with a 0.15% chance of occurring or 1 in 667. Where things start to get near their limits of realism is 4 standard deviations, 110.6 top speed, or 0.006% chance of occurring aka 1 in 15,700.

By setting 50% to the mean, and 100% to 4 standard deviations, you would end up with a pretty accurate reflection of the rarity of parts. Where this becomes tricky, and what I assume you are referencing NSG brought up, was that the different named parts have different means and SDs. The way pilots have categorized this for as long as we've known about it is basing the rarity on the named parts with the best drop chance for that stat. When you talk about the rarity of 8 engine Top speeds, you dont factor in the non QIDs because they don't have a chance to drop good top speeds, and the market naturally corrects to top speeds being more valuable because QIDs are a quest reward rather than standard loot. Where it becomes subjective is deciding what rarity top speed to pair with say a 95% rarity mass. There's never going to be a perfect answer to that because its going to come down to how many QIDs are being generated on your server. Despite that, the tables are not dynamic and do not shift over time. 95th % is 95th % forever

I recognize that script, first thing I did when I met NSG was to ask him about it, and this was his response. He assured me that the script generates the same results every time and would not be able to cause the shift.

Modified the loot tables, or modified the script so it kicks out different loot tables each quarter?