Observations about exalted imprints.

[Flitz]

I'll preface this with a quote from EHG_Mike;

"The best way to find a specific item is to imprint an item that's as close to the target item as possible."

In my experience that's not fully accurate, but bear in mind it's a mechanic that involves RNG heavily - it's entirely possible specific strokes of (un)luck and confirmation bias painted for me a picture that's quite different from reality and other people's experiences.

I was working specifically on double t7 imprints and paying attention only to double t7 prints. The sample size is more or less a week of nolifing and running echoes as quickly as I can. Here's the 'rules' I've observed:

• A printed item will differ in 1-3 affixes from the imprint. I wouldn't discount 0 and 4 being a possibility, but I can't recall any such cases, and it seems to most commonly tend towards 2.
  
• Exalted affixes appear to get changed to something else more commonly than low tier ones. When two affixes I wanted were both imprinted as t7, at least one of them would always get changed.
  
• Rare affixes are more likely to get changed. When imprinting a helmet with two skill prefixes, at least one of them would always get changed, even if they were low tier.
  
• If an affix gets changed, it's likely (but not guaranteed) to change to a 'similar' one. Fire penetration might become lightning one, % health might become flat health, dodge rating + crit avoidance might become just dodge rating. However, this is less likely to happen when both the affix and its similar alternatives are rare, as is the case with skill prefixes.
  
• Some specific affixes might be weighted differently than they are in natural drops. My only inkling for this is that I could never retain T7 movement speed from my imprinted boots, despite movement speed being a common affix.

Now, here's things I've been doing to rig things in my favour, seemingly with consistent success:

• Most important, is imprinting an item with my desired t7 position reversed. If I want two t7 prefixes, I imprint one with two t7 suffixes. This increases the odds at least one of my desired affixes being retained, and if one gets changed it's easier to fix because I can just use glyph of despair then rune of havoc (once if I'm lucky), instead of havoc, despair, havoc again.
  
• Secondly, prioritise common affixes for imprints, even though it might seem counter-intuitive. It's much easier to have your print retain common ones so you only need to craft the rare one on than it is the other way around.
  
• Lastly, one I only just started experimenting with, but it's showing some promise: if you need more than one rare affix, like on a helmet for example, consider an imprint with less than 4 affixes. I've seen some success in producing 3 affix prints this way so far, if 2 affix ones are also possible then this could save precious forging potential by bypassing the need to remove affixes that get rerolled, especially if you also want to utilize your seal slot for something useful or to safely change a suffix afterwards.

For a closing word, there's been a lot of discussion about imprints being nerfed, not just for uniques but for exalts too. It doesn't feel like it in my experience, at least when it comes to the quantity of prints, the RNG variance can be quite extreme. Last night I only had time for a quick one hour session and I got 3 double t7 sword prints, the day before I played for hours and got none. Now, have the prints gotten less accurate since the update? That one sounds more plausible, but I can't say for certain either.

Anyway, if you've been struggling with getting what you want from imprints, perhaps some of these steps will help you. Or perhaps it's all just pointless conjecture made on a too small sample size and I'm wasting my time writing this... that's RNG for you.

[Redeemer]

This should all be resolved by today's patch. What EHG_Mike stated is accurate, but the system wasnt reflecting.