Bungie was back this week with their This Week At Bungie update, going into more detail about Trials of Osiris, plus the further experiments they are planning for this weekend. The Flawless pool is no more, plus we got news of the future of Freelance, plus a nice little deep dive into the stats and analysis of how Trials is doing.
TRIALS MATCHMAKING CHANGES
The last few weekends of Trials of Osiris have been a bit quieter in terms of changes — this was by design! We wanted to give players a few weekends of steady play before making any more significant changes or testing out another Labs weekend. This gave our teams some great data for how the Trials changes are doing, how players are engaging with the activity, and what we could prioritize changing in the future.
A quick recap of stats from the last few weeks shows everything to be steady too:
Matchmaking times are consistent with past weeks.
Blowout rates dropped slightly week-over-week. This is consistent with lower-skilled players not playing as much.
Average hours played per player remains consistent across skill levels.
Overall player counts went down each week. But that is expected during this portion of the Season and is consistent with past Trials and Iron Banner performance.
We still don’t see any evidence of widespread Flawless card resets to avoid the Flawless pool — less than 0.5% of cards played were reset while still Flawless and above three wins. Even still, the impetus to do this should be removed starting this week — more on that below.
Now, we look to Freelance. Last weekend, we gave players a strictly “solo-queue” option for Trials. We entered the weekend with some assumptions on how the population could split, what might happen to matchmaking times, and a few questions on how to offer Freelance options in the future. Here’s a quick recap from the team on how things went, what we learned, and what we’re thinking about for the future.
So, this past week we opened the Freelance queue so that solo players could consistently play against other groups of solos. How’d that feel?
We had over a million games played in each playlist this weekend, with slightly more games played in the base Trials of Osiris playlist. Average hours played for the lower skilled players is up quite a bit over the past weeks, while the average to top-end has almost the same hours played. Interestingly, players who played in a fireteam some of the time and solo some of the time played almost three times as long this weekend as players who either played exclusively solo or in a fireteam.
244,000 Freelance exclusive players
120,000 fireteam exclusive players
220,000 played both solo and in a fireteam
25.5% of players went Flawless — lower than any revamp week — likely fallout of matches being fairer. If it doesn’t drop below 20% for a weekend, the Lighthouse should still feel attainable.
Average matchmaking times were consistently around 50 seconds for full fireteams and solos, and regularly were about 100 seconds for duos. The Sunday Flawless pool adds about 10 seconds to Flawless player matchmaking, and we are seeing some worrying spikes in matchmaking times for some of the smaller subpools post-Flawless pool split.
The overall blowout rate (games that ended with a score of 5-0) was almost as low as our best post-revamp week, when the Flawless pool went active Friday afternoon, at 27.8%. This was powered by the significant number of fireteam vs fireteam matches and solos vs solos matches, both of which had 25% blowout rates.
Blue – 10/5/2021
Orange – 10/12/20201
Match outcomes this week: Freelance / Base
5-0 Games: 25.1% / 30.2%
5-1 Games: 26.1% / 25.7%
5-2 Games: 20.4% / 18.5%
5-3 Games: 15.6% / 13.9%
5-4 Games: 12.4% / 11.4%
One thing we aren’t happy with is the team balance, and we’ve noticed that how the solo playlist isn’t doing a great job balancing two teams when there aren’t any fireteam mismatches. For Trials, it’s currently just picking teams randomly. We are going to be looking very carefully at how teams are selected over the next few weeks, and hopefully have a better balance the next time Freelance rolls around. This is one of the downsides of not using skill-based matchmaking in Trials: in situations where skill would be helpful – to balance two teams of solo players – it isn’t even a data point the system can look at. Speaking of which…
We liked the overall outcome and vibe of the Freelance queue, and both playlists seemed to have a viable amount of players. We are looking at running Freelance occasionally for the remainder of Season 15 but are still trying to determine what feels right. We will let you know!
Long term, we are exploring some single-playlist solutions: preferential matching of full fireteams against other full fireteams, duo + solo against another duo + solo. We don’t have a specific release target for this yet, but it will not be prior to Season 16.
FLAWLESS POOL CHANGES
The Flawless pool (which we have been enabling Sunday at reset time) has some very positive benefits. It raises overall match quality (blowout rates, etc.) significantly, and it raises the overall win rates of solo players. However, it has some drastic downsides: for players who struggle and manage to just barely go Flawless, it is a steep cliff to then only play against players who have already gone Flawless that week. It also incentivizes behavior that is at odds with the mode and reward systems, like resetting a Flawless six-win ticket to keep getting easier matches (but worse rewards). It also creates social friction where players need to decide between playing with friends or playing for rewards. Lastly, the timing — 10 AM Pacific on Sundays— is awkward if you live anywhere but the Americas.
Starting tomorrow, we are rolling out a new system and leaning into weekly performance metrics. Whenever you match in Trials, in addition to trying to match with people with the same number of card wins, you will also attempt to match with people who have roughly the same number of overall wins for the weekend. If you can’t find anyone to match against, it will eventually expand out to find you a less similar match.
For the PvP gods who go Flawless 10x a weekend, they will end up matching with other players who have (for example) 70+ wins on their cards, no matter how they got them. For the average team who gets a lucky Flawless in on Friday or Saturday on their first card, well, they will be matching against other players with around seven wins, no matter how they got them. We are hoping this will give us many of the benefits of the Flawless pool system (somewhat fairer matches) without the significant downsides. Specifically, there should be no major inflection point where everything is smooth prior to a win, and then everything is overly challenging after.
While this allows us to remove the Flawless pool, it does have a potential downside: longer matchmaking times. We think there is a reasonable tradeoff here between match quality and matchmaking time. Average Trials matchmaking times currently sit at around 50 seconds. Over the next few weekends, we are going to be looking at how much matchmaking time goes up, and how it affects the match quality.
The outcomes of the previous non-Labs week were:
Game outcomes: Overall % / Fireteam vs Fireteam / All Solos
5-0 games: 28.4% / 22.8% / 27.6%
5-1 games: 25.6% / 25.6% / 25.4%
5-2 games: 18.6% / 20.2% / 19.0%
5-3 games: 14.7% / 16.8% / 15.4%
5-4 games: 12.3% / 14.5% / 12.6%
In an ideal world, we want to get each of those numbers as close to 20% as possible, but we would rather see more 5-4/5-3 outcomes than 5-0/5-1.
As always, we are going to be listening closely to your feedback, as well as looking at aggregated stats to continue tuning matchmaking for Trials to strive for the best experience possible.
Let me know in the comments what you think of the latest Destiny 2 News about Trials.