These are master writ final statistics from Dragon Bones update, representing one very good crafter doing all six writs every day over three years. That is over 1100 days and over 7000 writs made.
This test nails down definite results for writ drop rates. It gives a good estimate for voucher drop rates (but with some margin of error).Definite master writ statistics: When doing all six writs, there is 72% chance to get one master writ. On average a crafter earns 7..9 vouchers each day.Click To Tweet
Let me first show the validity of both results:
The test run was 42 days long with 28 actual crafters. I divided the 42 days to three sets and calculated individual drop rates from each set. Then i calculated how much they deviate from total average (100% meaning total average). The idea is that while a test size is too small the results vary wildly. But the longer the test series is, more stable the results become. Dividing to smaller subsets and calculating them individually should show how far we have advanced in reliability.
The graph shows that the three blue subresults from writs are withing 5% of each other. This is usually a sign that the results are very reliable. Any RNG-based statistics i have calculated keep wobbling within 5%. This test run also fulfill my thumb of rule for statistically valid run: there has to be at least 100 drops of each item. Well, provisioning fails this slightly.
But three red subresults from vouchers still vary over 20% and this is a sign that the result is not yet valid. Any of those three results can be the correct one
In a nutshell: writ drop rate results are withing 5% and voucher results are withing 25% error margin. Getting the voucher results more reliable is too big effort.
- 1176 days crafting 6 writs each day on a simulated very good crafter.
- 7056 writs done.
- 842 master writs dropped.
- 9676 total vouchers earned.
- See the table for per-profession results.
Avg/month shows what a crafter can earn in 28-day long period (i know, not exactly a month).
Avg/charday shows daily averages for one character:
- There is 72% chance to get one master writ (from any profession).
- On average a crafter earns 8.2 vouchers each day. But because of the uncertainty, it could as well be 7 or 9 vouchers.
- See the table for per-profession results.
E.g. a blacksmith master writs have 10% drop chance. On average blacksmith master writs bring 2.3 vouchers each day.
The graphs don’t really bring in any new information. But they show how stable the drops are over time. The graphs show daily drops from 28 characters. The blue graph shows daily amount of writs. It’s pretty stable overall. The red graph shows voucher drops. It’s very clear that vouchers drops vary wildly. Even the trend line is slightly tilted.
To get reliable voucher results, the test run has to be at least three times longer (or even longer than that), meaning ten years of cratfting on single character. Even with 28 simulated characters it’s just too much work to do. Also, Summerset Isles update arrives within two months making such test run impossible.
I run the test with two accounts, 14 characters each. First account (or) has 9 traits on all except two characters have still few item slots with 8 traits. They have diligently done all crafting achievements and all know at least 10 epic motifs and epic recipes. The second account (ar) has one 9 trait on all while majority is halfway at 9 traits. They have all at least 10 epic motifs but are weak on recipes.
These numbers differ from the “results” so that each day has 28 characters and 6*28=168 writs every day. The “results” are simulated so that simulated time period is 28 times the real period, i.e. 28*42 days = 1176 days. Of course daily numbers are divided by 28 to get the averages for one character. One day with 28 characers is also nicely close to month, so that each real day is close to simulated month. This gets even better when the juniors at 15th slot character evolve enough. Then one day is equivalent to 30 day “month”.
Avg/day shows total daily averages from 28 characters.
Account Totals show totals divided to accounts. Usually the first account gets a little better numbers, but the gap is closing.
Account acc% shows the same as in Account Totals but as percentages of total values. If everything were equal, numbers should be 50%/50%. There is significant difference in provisioning. The second account characters have learned just few epic recipes. I need to fix them.
Account Avg/day shows daily averages per day per account, i.e. what 14 characters earn daily.
Character Avg/day shows daily averages per character per account.
I no longer gather results individually per character. The account is lowest level. I run writs on each character and move everything significant to bank. The i simply counted the master writs from bank afterwards every day.