Yesterday I disclosed the ratios of resonator drops depending on the portal level hacked. Today I would like to follow this up with the ratios for XMP bursters. The biggest difference is the likelihood to get a lvl+2 XMP compared to a lvl+2 resonator. You have 2 1/2 times higher chance (per item) to get a higher level one.
My proposed ratio is 3:12:4:1 which leads to x/20 or multiples of 5%. I have included the summary data for the 4000+ Glyph hack data below.
| Level | Weight | Theory | Observed | Observed Bonus | Zero Second |
| lvl-1 | 3 | 15% | 15.5% | 15.3% | 15.6% |
| lvl | 12 | 60% | 59.4% | 59.4% | 59.1% |
| lvl+1 | 4 | 20% | 20.0% | 20.4% | 19.9% |
| lvl+2 | 1 | 5% | 5.0% | 4.9% | 5.4% |
| Resonators | 3288 | 7641 | 2334 |
Other data sets I looked at:
- A lvl1/lvl6 small dataset from myself. It is inside the error margin.
- An even older set with >50K datapoints. This could also be explained with an alternative 4:15:5:1 ratio. It would imply that the lvl+2 data was boosted at some time between 2014 and early 2015 by 25%. This would be in line with a boost of the lvl+2 resonator drop rate by 100% at the same time.
The difference in relative level distribution is shown in the graph. lvl-1 is nearly identical for resonators and XMP. But XMP are boosted for lvl+1 and lvl+2 compared to resonators.
Table 1 and 2 are the normal hack data from Philip Kloppers
| Portal | lvl -1 | lvl | lvl +1 | lvl +2 |
| 1 | na | 689 | 177 | 40 |
| 2 | 133 | 522 | 174 | 40 |
| 3 | 114 | 489 | 165 | 41 |
| 4 | 182 | 710 | 244 | 70 |
| 5 | 62 | 262 | 79 | 17 |
| 6 | 123 | 364 | 130 | 30 |
| 7 | 29 | 134 | 43 | na |
| 8 | 0 | 0 | na | na |
| Portal | lvl-1 | lvl | lvl+1 | lvl+2 |
| 1 | na | 76.0% | 19.5% | 4.4% |
| 2 | 15.3% | 60.1% | 20.0% | 4.6% |
| 3 | 14.1% | 60.4% | 20.4% | 5.1% |
| 4 | 15.1% | 58.9% | 20.2% | 5.8% |
| 5 | 14.8% | 62.4% | 18.8% | 4.0% |
| 6 | 19.0% | 56.3% | 20.1% | 4.6% |
| 7 | 14.1% | 65.0% | 20.9% | na |
Followed by the bonus hack data
| Portal | lvl -1 | lvl | lvl +1 | lvl +2 |
| 1 | na | 1092 | 276 | 76 |
| 2 | 233 | 943 | 304 | 72 |
| 3 | 235 | 1040 | 389 | 80 |
| 4 | 438 | 1635 | 559 | 137 |
| 5 | 146 | 567 | 181 | 53 |
| 6 | 258 | 916 | 321 | 81 |
| 7 | 76 | 379 | 150 | na |
| 8 | na | na |
| Portal | lvl-1 | lvl | lvl+1 | lvl+2 |
| 1 | na | 75.6% | 19.1% | 5.3% |
| 2 | 15.0% | 60.8% | 19.6% | 4.6% |
| 3 | 13.5% | 59.6% | 22.3% | 4.6% |
| 4 | 15.8% | 59.0% | 20.2% | 4.9% |
| 5 | 15.4% | 59.9% | 19.1% | 5.6% |
| 6 | 16.4% | 58.1% | 20.4% | 5.1% |
| 7 | 12.6% | 62.6% | 24.8% | na |
My own very restricted data
| Portal | lvl-1 | lvl | lvl+1 | lvl+2 |
| 1 | na | 99 | 30 | 13 |
| 6 | 15 | 88 | 27 | 4 |
| Portal | lvl-1 | lvl | lvl+1 | lvl+2 |
| 1 | na | 76% | 23% | 10% |
| 6 | 11% | 65% | 20% | 3% |
And the 2014 data from hack-log public
| Portal | Weight | Theory | A<>A | Percent | R<>R | Percent |
| -1 | 3 | 15.0% | 1989 | 15.6% | 5961 | 15.0% |
| 0 | 12 | 60.0% | 7720 | 60.5% | 23730 | 59.9% |
| 1 | 4 | 20.0% | 2564 | 20.1% | 8135 | 20.5% |
| 2 | 1 | 5.0% | 494 | 3.9% | 1794 | 4.5% |
And the zero second (negative time) single hack
| Level | Amount | Percent | Predicted |
| lvl2 | 363 | 15.6% | 15.0% |
| lvl3 | 1380 | 59.1% | 60.0% |
| lvl4 | 464 | 19.9% | 20.0% |
| lvl5 | 127 | 5.4% | 5.0% |
| all | 2334 | 100% | 100% |