Guild Wars 2

So, I want to put a foil hat on, and see if I can hand them out to others
I have done a few forgings to make Amalgamated Gemstones and have noticed a steep decline in how many you get, starting about a week ago.
The first week, I saw about 11.7 AG/forging (1 in 8.9 forgings give 25 instead of 10)
Now, it’s about 11.2 AG/forging (1 in 12.5 forgings give 25)

Now, I know what you’re thinking, because usually when people bring this sort of thing up their data looks like, “I forged this 5 times and only got 50 AG, shenanigans!”. However, I assure you my sample size is sufficient to have this discussion:
on 4/22 I reported 445 forgings gave 5,200 AG
After that, I did more forgings and the average peaked around 11.8 AG/forging. But I don’t have exact numbers for that, as I didn’t expect to need them. It included about 800 forgings.
5 days ago I got suspicious when my average plummeted, so I started recording seperate data: 416 forgings which gave 4,685 AG (11.26 AG/forging)
even worse, my results from today: 78 forgings gave 840 AG (10.77 AG/forging)
Total: 1,460 forgings gave 16,820 AG (11.52 AG/forging)

Anyone else observe this? Can we bring out the pitchforks?
I know this sample size could be bigger, but if you say this isn’t big enough to start this discussion, nothing will satisfy you. Believe me, this is big enough that the odds of just being unlucky is astronomically small.
Really, the only thing I’m annoyed about is that my initial data is obsolete; it’d be nice to get some confirmation so I could trash it.

While I wouldn’t count a 0.6 AG/forging a steep decrease in the drop rate, I don’t know if the difference is statistically significant or not (hard to tell when each forge can give more than 1 AG). So I can’t say that there hasn’t been a decrease in the drop rate.

If it is, it does show that ANet needs to do it’s computer simulations in order to determine with higher degree of accuracy if there is a problem or not.

If the true average is supposed to be 11.5 AG/forge, then your results I wouldn’t say are problematic: you just happened to hit the high end and the low end of the bell curve. But if it’s supposed to be 11.2, or 11.8 or higher or even lower, then there looks like there may be something up with the code.

Instead of looking at the AG/forge rate, look at the crit rate for getting 25. Assuming your math is correct, your data shows a drop from better than 1:9 to less than 1:12 (or from 11% chance down to 8%). That’s a pretty steep drop to be explained simply by luck, with over 400 trials with the good luck and another 400 with the bad.

Although that’s not enough data to say for certain that something has changed, it’s imo more than enough to ask Probably John Smith to look into it. You can try summoning him via Reddit, the in-game /bug reporting tool, or whisper him here.

(And sure, it would be good if another person could confirm your results, but you have a long history here on the forums — I trust your data nearly as much as I trust my own (although tbh, I don’t even trust mine 100% ) .)

hmmmm, yeah, now that you mention it, reddit might be a better spot to get JS attention