Oracle Post-Phase Analysis

After a week of looking through posts, threads, tables of data and graphs, and oxymorons aside, it’s time for a bit of an Oracle retrospective. Here are some of our bigger takeaways from our most recent blitz phase.

Victory Points

As we expected, our VP heatmap clearly shows one of the more defining features of the Oracle Phase. In many regions, the overall VP spread was incredibly narrow—particularly at the top of the boards—and many of the regions went back and forth through the phase.

Notice how some regional winners didn’t even make it into the heat map? This is because the table lists only the top 20 overall VP winners. Some Alliances, however, were able to win their region without breaking the top 20 in overall VP. For the record, here are the regional winners again.

And while not all regions were as hotly contested as Stormlands and The Vale, there is no doubt that Cold Dark Kongnanigans and Kings of Winter fought hard against all comers to gain and maintain their winning positions.

Here is a finer-grain look at these VP growths put together by our data maesters. These cover VP growth over just the first 12 hours of Oracle. 

stormlands_hourly_vp.png

VP Calculations

Before we dive too much deeper into this blog, we want to address the concerns of those players who noticed inconsistencies in the calculation of VP at various points in the phase. Looking into it, we discovered two aspects to these inconsistencies. The first was that from time to time, the Passive VP calculations, run on 15 minute timers, seemed to “skip a beat” and not add in the appropriate amount of Passive VP. This omission, however, would automatically be corrected on the next cycle and Passive VP would effectively correct itself. We do not have an explanation for this, though we believe that there may be governing AvA code that we overlooked changing for the phase.

Additionally, we can confirm that some Alliances appeared at times to be awarded more VP than should have been theoretically possible. While we do not believe this impacted the overall leaderboard positions by the end of the phase, it is something we intend to look at closely ahead of our next blitz phase. 

Harmful & Helpful Actions

As for Harmful and Helpful Actions, at first glance, you may be excused for thinking that Oracle was less eventful than our usual phases.  

harmful.png

Far from being uneventful, however, Oracle was simply differently eventful. The Helpful and Harmful tables represent the actions only of the 10 overall Alliances in terms of VP. As you may assume, this pool coincides in regular phases generally with the Alliances giving and receiving the most actions. With the inclusion of a “Proximity VP” in Oracle, however, actions were aimed at the closest Alliances, which were constantly shifting--unlike the top of the leaderboard, which is the most common target in regular phases. As a result, the actions received and given were more diffuse. As well, like the heatmap above, many of the most active Alliances regionally, simply didn’t make it into the pool of top 10 VP Winners overall. Finally, the change we made to Spy actions that allowed fewer spy attacks on a single Alliance for them to gain full “intel,” meant that Spy actions were less concentrated.

Proximity VP & Active to Passive VP Balance

We were pleased to see a generally positive response to Proximity VP in Oracle. Many players found it offered a refreshing approach to AvA provided their Alliance more strategic choices throughout the phase. Despite a number of requests to include it in regular phases, however, we do not feel that Proximity VP by itself would be as successful outside of Oracle. 

Proximity VP was effective in Oracle because it was one of a number of other connected changes. These other changes included disabling of Silver contributions, shorter VP timers and a lack of negative consequences on losing actions. Outside the context of these changes, in fact, proximity VP may result in the largest of Alliances becoming even more dominant. So, while we intend to include proximity VP in future blitz phases--with experimental tweaks, of course--it will not be added to regular phases at this time.

Compared to the previous blitz phase, Rogue, we significantly increased the amount of Active to Passive VP in Oracle. Many players felt that this ratio was markedly improved--and the level of competitiveness depicted in the Hourly VP Growth Graphs (above), indicates that Active VP did play a bigger role in Oracle. Others, meanwhile, feel that Passive VP was still too high relative to Active. We are continuing to analyze the Oracle data and will again tweak this balance for the next blitz phase.

XP in AvA

XP from AvA Actions was another popular Oracle component. As we had hoped, players appreciated the ability to level up, provide themselves with more GP for future AvAs and not have to take a break from leveling just to play AvA if they are a reincarnator.

It remains our intention to make AvA XP a permanent part of Game of Thrones Ascent. Our aim in premiering it in Oracle wasn’t to make it a one-off, blitz component, but to allow us to test its impact on our backend systems. Unfortunately, we discovered that our approach to calculating XP, despite doing well in our test environments, added undue stress on our systems once scaled up in the real world. 

So, currently we are developing a new approach to providing XP in AvA; one that that will not pose a risk to back-end performance. While a solution will not be ready for the Traitor Phase in December, we are confident that AvA XP will roll out in early 2018. This is our highest AvA priority.

GP Cap

Capping the Garrison Points for all players was unpopular with a number of Oracle participants who felt that GP should adjust to and reflect their character level, like it does in regular phases. Our aim with this restriction was to “level the playing field” a bit, to help make the blitz phases more competitive for more players. We recognize, however, that many established players understandably feel they have earned their strategic advantage.  All's fair in love and war, some might say, especially in AvA. Truth be told, in addition to compensating for increased Sworn Sword cooldown timers, we raised the GP cap to 40 from the 20 used in the Rogue Phase to help balance earned strategic advantage with phase competitiveness for more players.  

In our next AvA blitz phase we will take a closer look at the GP question and try to achieve an even better balance of these two needs.

Superlative Titles

Like XP, Superlative Title rewards were a near-universally liked aspect of Oracle. It is also an aspect we are happy to continue into the foreseeable future (in fact, Smuggler already includes Superlative TItles as of this writing). As we’ve noted before, the purpose of these titles it to provide more opportunities to appropriately reward AvA participants for their efforts and skill. Like the second-rank regional titles, Superlative Titles offer an additional way to approach and win at AvA.

One question that has emerged, however, is the use of “per participant” versus “per alliance” Superlative Titles. We feel that including both offers a wider range of options and possibilities, though we will do our best to select Superlatives that do not automatically coincide with regional winners. This will keep Superlative Titles as winnable as possible by the full range of Alliances.

In Conclusion

We want to thank all Alliances for taking part in Oracle. Despite its difficult start—and thank you again for your patience with that—the phase ended well. Many players reported that they enjoyed the blitz phase, and provided constructive and positive feedback on the experimental changes. We will offer more blitz phases in the future, of course. We look forward to any suggestion and ideas you have to make them more exciting, competitive and innovative.