As a service to my fellow forum members, i decided to collect some empirical data on talents to help determine their level of randomness.
I decided to follow
@CRABNUTZ 's advice and continue to roll until I received the 5/5 talent I was looking for (in this case 5/5 Missile Damage).
I repeated this experiment 5 times (feel free to PayPal me cash if you would like additional data sets).
It took me 299 rolls in total to obtain the five talents. This is an average of 60 rolls per talent or 3,000 gold under the current 50 gold refresh price. This is slightly less than
@CRABNUTZ 's estimate of 3,500 gold. I received the perfect talent in:
Try 1: 68 rolls
Try 2: 30 rolls
Try 3: 32 rolls
Try 4: 82 rolls
Try 5: 87 rolls
My theory before I started was that talents received earlier in a string of rolls were more likely to be lower level and as the rolls continued to be made, the chances of obtaining a perfect talents would increase.
For purposes of quantifying the data, I grouped rolls into sets of 10 (first ten rolls, then second ten, etc. ). Here is the data I collected:
Rolls 1-10 (50 rolls)
1/5 Talent 30%
2/5 Talent 22%
3/5 Talent 24%
4/5 Talent 20%
5/5 Talent 4%
Rolls 11-21 (50 rolls)
1/5 Talent 24%
2/5 Talent 16%
3/5 Talent 32%
4/5 Talent 22%
5/5 Talent 6%
Rolls 21-30 (50 rolls)
1/5 Talent 28%
2/5 Talent 28%
3/5 Talent 14%
4/5 Talent 20%
5/5 Talent 10%
Rolls 31-40 (32 rolls)
1/5 Talent 25%
2/5 Talent 22%
3/5 Talent 28%
4/5 Talent 12.5%
5/5 Talent 12.5%
Rolls 41-50 (30 rolls)
1/5 Talent 7%
2/5 Talent 27%
3/5 Talent 33%
4/5 Talent 23%
5/5 Talent 10%
Rolls 51-60 (30 rolls)
1/5 Talent 27%
2/5 Talent 37%
3/5 Talent 17%
4/5 Talent 10%
5/5 Talent 10%
Rolls 61-70 (20 rolls)
1/5 Talent 11%
2/5 Talent 18%
3/5 Talent 14%
4/5 Talent 46%
5/5 Talent 11%
Rolls 71-80 (20 rolls)
1/5 Talent 25%
2/5 Talent 5%
3/5 Talent 20%
4/5 Talent 20%
5/5 Talent 30%
Rolls 81-90 (9 rolls)
1/5 Talent 0%
2/5 Talent 11%
3/5 Talent 22%
4/5 Talent 0%
5/5 Talent 66%
Even though the above data is from a lot of rolls, the sample size is still rather small to draw foregone conclusions. Also, the fact that I stopped gathering data once I hit a specific 5/5 Talent skews the data as well. And, the sample sizes for the latter groups of 10 are small and a couple of different results would paint a different picture.
Now, with those disclosures out of the way, as I look at the data, there certainly seems to be an increased likelihood of higher level talents as you progress through the rolls with a significant shift around the 60 roll mark.
The actual talents offered appear to be fairly random (i.e RPG Health, Firebomb damage, etc.).
That being said, the other two premium talents that I would want (Flash and Firebomb Damage) were received only once each during these rolls. I received 5/5 Flash on Try 2 on my 80th roll (#82 was 5/5 Missile). I received 5/5 Firebomb on Try 5 on my 56th roll (#87 was 5/5 Missile).
I hope this helps. If i make another talent run I might update my results.
