Overwatch: Season 2 Gets Rid Of Sudden Death

Gather ye round for details of the Overwatch [official site] competitive season 2 changes, direct from the maw of Jeff Kaplan. The changes currently include the abolishing of sudden death rounds, a new skill tier system and more:

Too long; didn’t maw?

So the prompt to get rid of sudden death rounds to deal with tied results was because players really didn’t like the way losing (and possibly winning) felt with those matches. Instead Blizzard will be switching assault, hybrid and payload maps to use a time bank system where how fast you end one round impacts the time you have on the next. The video has the full details of how that works and how it’s different from current time bank aspects you’ll find in the game. Control maps are sticking with the season 1 method where you win three out of five matches, by the way.

The trade off here is that you’ll sometimes get a match which ends in a tie rather than always having a winner, but Kaplan’s explanation is that ditching unsatisfying sudden death will be worth the [rare] tied match.

The other big change relates to how players internalise or identify with their skill rating. The previous season’s skill rating used a 1-100 system to grade players and players would increase or decrease along that, sometimes by fractions of a point.

Listening to Kaplan it sounds like the 1-100 wasn’t representing linear progression, but the normal distribution curve you tend to get in games like this where most people are clumped in the middle and then you get some super-skilled or super-awful outliers at either extreme. Kaplan says that with a skill rank of 60, players tended to respond as if this wasn’t a particularly exciting thing, maybe seeing it as a grade D on a school paper when in fact “you were in the top 6% of all Overwatch players”.

He also says that people would invest incredibly hard in that single number, getting really upset if it dropped a bit after a bad game/bad few games.

In response Blizzard are switching to a new system. You still get a skill ranking but it’s on a scale which deals in thousands and not 1-100 and so you’ll go up or down whole numbers. But the important part is that there are new tiers every 500 skill rating so players can identify as a “gold” or “diamond” or “bronze” and have that be more of a reliable/stable competitive identity. There’s essentially a bit more wiggle room so that losses don’t have as much capacity to affect the skill identity you have.

The tiers are the usual precious metals and mastery nods: bronze, silver, gold, platinum, diamond, master, grand master. For everything below master tier, once you’ve achieved that tier you never drop below it (i.e. there’s a high water mark aspect to tiering). For the high ranks you can drop down as far as diamond if you have a nightmare week/month/season.

You also start to suffer from skill rating decay. After seven days of inactivity players in diamond, master and grand master will lose 50 points per 24 hours. If you take too long a holiday you’re at the butt-end of diamond and will need to clamber back. Rewards at the end of the season are based on the highest tier you achieved.

Other topics touched on in the video include that Blizzard are narrowing the difference you’re allowed in skill ranking when grouping up with other players, changing the competitive points system and so on.

The changes will be up on the public test realm (PTR) if you want to give them a go, and more exact patch notes will probably arrive when the team has had a chance to process the responses to the changes from said PTR.


  1. Meat Circus says:

    Thank the Dark Lord.

    Look, I love Overwatch, but we can all accept CP1 was a poorly designed, unfair and mostly thankless salty mess. Hence virtually nothing surviving intact to CP2.

    I have to admit I’m slightly surprised Blizzard were happy shipping it in that state.

    • airtekh says:

      It was a tough decision.

      They said they’d rather people play an incomplete competitive mode, rather than no competitive at all.

  2. Premium User Badge

    Oakreef says:

    That mercy image looks like it should have a “come play, my lord!” caption over it.

  3. eeguest says:

    Is it possible to get skill rating up even when losing a match? (Because you played so good but your team was bad as whole)

    • Beefenstein says:

      In placement matches yes: your placement depended on how well you played, not winning or losing.

      Outside of those placement matches (first 10) no. If you are good for your team you will help them win until you find your level where you win and lose at the same frequency. This is then your ranking which, theoretically, will only fluctuate due to randomness unless you personally get better or worse. In practice I suspect this system works pretty well considering the fluctuating effects of you or other people in the game having a bad day/being drunk/playing a character they don’t usually and are very good or bad at in comparison, playing with heroes that are more or less effective in that situation etc. Oh, and changes to the meta due to hero buffs and nerfs will change ranking too — if defence heroes were all buffed to God-tier it would rearrange placement.

      • Kitsunin says:

        Actually, while the effect appears to be much lesser than in Placement (I went 3W/7L and got placed at rank 55) individual performance offsets the rank change somewhat.

        That is, while you’ll never ever gain rank from a loss, you will lose less/gain more if your “performance” is better. Unfortunately this only seems to relate to scoreboard stats relative to the total average, so some characters (namely: Mercy) always receive much positive performance-based rank because their numbers won’t be as high as others’.

        • Haxton Fale says:

          What you said seems to run contrary to what is observed. Top teams had their dedicated healers place some ten SR below the rest, and playing mainly Mercy I see my SR change not by an awful lot. If anything, good Mercy on a mediocre team that benefits the most from having her will skew the averages against a good Mercy on a good team that does not need her as much.

          It’s hard to track Mercy’s contributions to team. The way OW does it is clunky, awkward, and takes away a lot from the picture, especially since nine times out of ten Lucio will beat her in raw amount healed, while scoring medals in other categories. At the same time I can’t really think of a better stat to track, perhaps the percentage of team damage healed?

          I’m generally not too fond of the current SR system and the fact that you lose rank on a loss regardless of your performance, which might punish most heroes who can’t carry a team to victory – all it takes is a single ally who is not doing their job properly, everyone else might perform admirably, but a loss is a loss and the rating goes down. And to say nothing of curb-stomp matches which seemed quite prevalent in my solo-queue career.

          • kalzekdor says:

            A good stat to track for Mercy would be assists, both Offensive and Defensive. Because of the focus on healing done, everybody neglects the Staff’s alt fire. I’ve dealt with too many Mercys who sit there with a healing beam on me while I’m at full health, for several minutes. With any halfway decent team, if all you do as Mercy is heal, the team would be better off with a Lucio. Well placed damage boosts can make a huge difference.

          • Haxton Fale says:

            A good starting point would be to track damage boosted in the first place. Assists are one thing, and a useful pair of stats no less, but I would also like to see the general efficacy of the damage boost recognised by the game in the form of a number. After all, everyone has a separate Damage Dealt stat, next to Kills and Objective Kills.

            But yes, using Offensive Assists combined with Healing would give you a pretty decent view of a Mercy’s usefulness.

          • kalzekdor says:

            I would actually like to see a combined overall stat that takes into account all your contributions to the team, weighted differently for each character. Besides the basic things, e.g., damage done, eliminations, objective time, etc., it would also give bonus points for certain actions based on the character’s role. Mercy would get points for Offensive Assists, Resurrections, and maybe a small multiplier to points gained from healing low health allies. Snipers could get bonus points for taking out Support characters. Pharah and Junkrat could get bonus points for taking out Bastions or Turrets. Tanks could get a small amount of points for blocking damage, and bonus points for holding an objective. Tracer could get bonus points for finishing off enemies. McCree could get points for headshots and stuns. Hanzo and Widowmaker could get points for Recon Assists. Etc., etc., etc.

            Basically, instead of focusing on the same basic stats for all characters and roles, reward each character differently based on the unique contributions they can make to the team. Why bother having distinct classes and roles if you track performance irrespective of those roles?

  4. Banks says:

    The competitive season has also evidenced that the game is horribly unbalanced and most characters need a major rework. Everyone is complaining about Genji and saying he is OP, but the truth is that he is the only hero that rewards decent skill.

    In competitive, no one chooses defense heroes, they are badly thought and they require zero dexterity to make them fun for skilled players. They aren’t very effective either. Widow is fine I guess, if you can tolerate everyone cursing against you for picking her.

    The same goes for support and tank heroes. Ana an Zen are fun, but Mercy and Lucio are easy as hell and yet they’re the most effective. Symmetra is just a mess. The tanks are in a better place because they are situational, they are mostly fun and very helpful if chosen rightly. However, I would love if they added a couple more tanks that are hard to master.

    Which honestly is my problem with the game at it’s current state, the game desperately needs more heroes with high skill ceiling.

    • Beefenstein says:

      “Ana an Zen are fun, but Mercy and Lucio are easy as hell and yet they’re the most effective.”

      Current meta is Zen-Lucio. Zen nerf might alter this. Zen, right now, is a beast.

      I do agree that Genji is the highest skill-ceiling character at the moment. There are other characters with appreciable skill ceilings too, though.

      • Tenerim says:

        While I also agree that Genji is the highest skill ceiling character right now, I strongly think he should get a nerf. There are a few things that I would like to see tried at least, maybe they are too harsh. Namely: Add a 2 or 3 second cool-down on his double jump or make it so his ultimate has the range of a normal melee attack instead of from the peak of his double jump all the way to the ground or make it so that getting a kill during his ultimate does not refund cool-down on his shift ability. I’d like to see if any of those three changes would put a damper on how effective Genji is. Right now when a Genji pops his ultimate at least all the healers go down and possibly the entire team if enough distraction is happening.

        • Winstons says:

          Serious lollers at these suggestions.

        • Evil Pancakes says:

          I agree with you on Genji being too effective at what he does, while also being the highest skill cap character in the game. Don’t agree with most of you nerf suggestions though. Personally, I think he really just needs a nerf to his ult. Right now, his ult is just too good. It’s not even that difficult to get 3-4 or more kills with the one ult, and he’s really tricky to stop with how slippery he is. Compare that to other high damage potential ults, like rocket barrage from Pharah. Pharah is stationary in the sky which makes her a massive target and as a result almost always goes down when using it. Or compare it with Tracer. Her ult charges only slightly faster than Genji and she fits the same player role. Yet her ultimate is much less potent. Hell, it isn’t even enough to kill most tanks in one shot. Phara and Tracer’s ults are best used when paired with another crowd control ult to really get 3-4 kills out of an ult. Genji can just solo that, so why use anyone else?
          So what I would like to see is just tone down the range of Genjis ult and maybe the duration, say down to 5-6 seconds from 8. It’ll probably still be enough to kill most high value targets, i.e. healers, but not enough to kill over half the enemy team. Not without support anyway.

  5. geldonyetich says:

    I was hoping the “removal of sudden death” referred to not getting one-shot by snipers. No such luck.

    • Kitsunin says:

      Because they already did, by making the snipers unviable in competitive. Unless you count McCree.

  6. Kitsunin says:

    Literally a typo, the comment system ate the “less” after “much” :(

    Still, those are nice specifics.

  7. Premium User Badge

    keithzg says:

    Instead Blizzard will be switching assault, hybrid and payload maps to use a time bank system where how fast you end one round impacts the time you have on the next.

    How innovative, they’re finally doing what Unreal Tournament did in 1999!

    (I love playing the game, but it’s painfully clear at times that the designers are bizarrely unaware of the existence of non-Blizzard games at times.)