Difference between revisions of "Experimentation Source 11 (Source)"

When there are two different formulas for the same experiment line, there are actually 2 experiment percentages internally. But you are only shown an average of those 2.

Each time you experiment one point, a separate roll is made for both of those percentages. But the result is not applied directly, instead a factor is used which tries to compensate for one being lower. For example if one is 20% and the other is 30%, you will see 25%. Then the roll for the first one is multiplied by 1.2, while the second is multiplied by 0.8. This will bring the factors closer and closer together until you reach the the first maximum.

Here is where the trouble starts. The system will continue to use this compensation even though you have already maxed one percentage. So you will get less and less on the advancing percentage the further you get from the the maxed one as it is trying pointlessly to divert more to the maxed one. I'm not sure if it's supposed to work this way, or if it was an oversight by the developers.

The bottom line is that it's always a bad idea to have percentages too far apart on the same experiment line.

One more thing. Since there are 2 rolls, you can get different successes on them. The status you see is the lowest of the two. That is why you see far less amazing successes on the final buff pack than simple components. You will only see an amazing success if both gets an amazing one. The funny thing is that it seems to continue to roll for the one that is maxed out. I only got a 'good success' reported after maxing out the duration in my tests. But the power still increased as if I had gotten a normal great result.