Day 3 results:
Test Subject | Footfall | Repeat | # in 5 days | Days since last visit | Test Subject | Footfall | Repeat | # in 5 days | Days since last visit |
---|---|---|---|---|---|---|---|---|---|
1 | 84 | no | 1 | never | 11 | 33 | yes | 2 | 2 |
2 | 84 | no | 1 | never | 12 | 21 | yes | 3 | 1 |
3 | 84 | no | 1 | never | 13 | 33 | yes | 2 | 2 |
4 | 84 | no | 1 | never | 14 | 21 | yes | 3 | 1 |
5 | 84 | no | 1 | never | 15 | 33 | yes | 2 | 2 |
6 | 84 | no | 1 | never | 16 | 21 | yes | 3 | 1 |
7 | 84 | no | 1 | never | 17 | 33 | yes | 2 | 2 |
8 | 84 | no | 1 | never | 18 | 18 | yes | 3 | 1 |
9 | 72 | no | 1 | never | 19 | 28 | yes | 2 | 2 |
10 | 72 | no | 1 | never | 20 | 18 | yes | 3 | 1 |
Some notes about this round of tests:
First, due to the complexities of scheduling, the first ten visitors (that had never visited before) had to happen a bit earlier than I would have preferred, so I don’t think there had been enough time for the ‘visits in the last 24 hour hours’ counter to go all the way to zero. I think it can be safely assumed at this point that this is an important counter.
Second, when I first unsealed the area and went to check the beacon before the testing, there was 72 there, so I suppose some random first-visitor may have been lured to my test area by the portal icon in the compass (@james , would be fantastic if we could opt out of that in the portal UI ) and ended up stepping on the ‘ceiling’ of the test beacon (it’s buried under a layer of natural soil but that soil is still part of the beacon), no idea when this visitor showed up.
Then, since I’m a crazy person, I decided to alternate the order in which the friends helping with the second half of the test would enter the area. 5 had been away from the beacon for 2 days and generated more (so it seems to go up the longer a particular character stays away) and 5 had been visiting daily.
I decided to alternate because at this point it seems clear that there’s a multiplier based on the number of visits in the last 24 hours.
The ‘tiers’ for the top footfall for new visitors so far observed were:
Post-211 | Pre-211 | % | in last 24h: |
---|---|---|---|
120 | 60 | 200% | 0 to 5 |
96 | 60 | 160% | 6 to 10 |
84 | 60 | 140% | 10 to 20 |
72 | 60 | 120% | 20 to 30+ |
Since most tests so far were of about 20 visitors a day, that is all that could be inferred so far.
Now, we can extrapolate that those multipliers would also be applied on top of the ‘reduced’ value for repeated visits in the 5 day period. That is why I decided to alternate between the second-visit and third-visit test subjects, as you can see in the first table above they were giving consistent values, and then when the threshold for the next tier of ‘visits in the last 24 hours’ was hit, both numbers decreased a little.
This round of testing also indicated that the footfall generation of a character goes up the longer the character stays away without visiting.
Interesting results so far, and I think some solid hypotheses can be extrapolated so far. I’m still curious to see what will happen to the repeated visitors after five days of testing. We’ll see.
Incidentally so far the running total indicates that (for isolated beacons such as this one) there was a definite improvement over the pre-211 values.
Post-211 | Pre-211 | Pre-211, 24d | |
---|---|---|---|
Day 1 | 1680 | 1200 | 1200 |
Day 2 | 1308 | 1320 | 720 |
Day 3 | 1075 | 1200 | 600 |
Total | 2988 | 2520 | 1920 |
As usual, hit me up with any suggestions for things to test or look out for.
Edit 1:
However, in the hypothesis that the ‘5 day’ timer requires visitors to stay away for the full 5 days before they can generate the ‘full’ amount again, the running total will end up indicating that the current system is better for hardly-ever-visited beacons, with pre-211 being better over time for repeatedly visited beacons (such as popular shops, malls and portal hubs) and pre-211 with the 24 day cooldown being the worst by far. Server resets wiping the timer during the 24 day cooldowns would have evened out that side of the table by a little, but it should still lag behind the other two models for popular beacons.
Edit 2:
Some more maths:
Repeated visits do not decrease the payout after the first.
If we plug the payout for the visitor that came every day to the multiplier table we get:
Base Payout | Bonus | Final Payout |
---|---|---|
15 | 200% | 30 |
15 | 160% | 24 |
15 | 140% | 21 |
15 | 120% | 18 |
All of those values have been seen for the daily visitor, some during day 2, some during day 3. So for repeated visitors, the payment floor seems to be 25% of the pre-211 base value, in this case 15c, and then the unpopularity multiplier is applied.
In the same way, we can infer that the increase from the floor of 15c for the visitor that skipped a day was to 24c and the values rounded down:
Base Payout | Bonus | Final Payout |
---|---|---|
24 | 120% | 28.8 |
24 | 140% | 33.6 |
Nice and round numbers:
Base (pre-211) | Base now | % | Days Away |
---|---|---|---|
60 | 24 | 40% | 2 |
60 | 15 | 25% | 1 |