The secret art of rubberbanding

Unknown · March 2014

Hello fellow MPQers,

I was hoping someone could enlighten me to the tricks of this system and how I can use them to my advantage.

Reason I ask is that I was top of the Heroic Jugger event last night for probably around 7 hours, always fending off people who were closing in on my points. Then in the last hour and a half, the top ten completely changed. I was still on top, but the other 2 people who were chasing me were now in 3rd and 6th respectively with people who I hadn't even noticed showing up from nowhere. I imagined that they waited until the last moment to enter then took the battles for huge points, which I still thought was weird because I had been playing for a long time climbing the ranks and even when I got to the top ten, my larger points dried up meaning I had to climb slower. I never instantly went in at number #2.

Anyway needless to say that within another 3 battles, although I was still winning battles for 90 - 120 points, I was knocked down to 7th. With an hour to go and being in the UK (3am) I was in no mood to continue only to loose out to an exploit, so I quite, waking in the morning to find I'd ranked top 50...... Dire. Now to me that was like the game saying, "Thanks for all your devotion and hard work, YOU JUST GOT SCREWED".

So anyway, using that old chestnut of "If you can't beat them, join them", which is incredibly apt in this instance, I was wondering if people could impart their knowledge with me on how to use rubberbanding to my advantage, the point refresh system, bracket entry and any other useful tidbits to put me on equal footing as those others.

I mean it's quite clear to me that if you don't have 8 115 lvl characters, then you are just not meant to win a X-Force or Invisible Woman, which frustrates me even more because a lot of the time these people already have full abilities on those characters (I'm looking at your EvilSonGoku, stop getting in my brackets dude). Evil, evil people. Plus I don't like pumping purple into just 3 people, that's never been my play style and I like finding cool matches throughout my roster. I mean right now I'm loving modern Hawkeye. Favourite move is Avoid. It goes like this

Juggernaut, 6 reds, headbutt on Hawkeye.
Hawkeye, 6 purples on the board, taps Thor on the shoulder, "Hey Thor, Juggernaut wanted to tell you something"
Hawkeye Avoids.

Honestly it cracks me up everytime.

Thanks in advance for any assistance.

Unknown · March 2014

Excellent post, I would like some feedback on this as well.

Arogntbastrd · March 2014

Rubber banding in pve is a little trickier because even if you wait, clearing the pve once won't be enough to get into first (at least that's been my experience) then there's the enemy scaling which gets in the way if you have a mid tier roster. It is nice though when you want a mid tier reward but don't want/ have the time to grind endlessly. Ideally you'd want to grind the stages down to 1 point right before the reset then wait again and repeat. That's how I've had success in longer pve

In pvp it's a little simpler, you just have to wait and time it right to where you can get to first as close to the end of the tournament as possible. This is easier if you have some high lvl characters bc you'll get less retaliations along the way.

The ultra elite players even go as far as to play, shield, find high point targets then take them down and reshield. I believe this was the go to strategy in the pay to play tournaments.

I'd suggest not sweating trying to get first now. Decide what your realistic finishing position is and just try to get there consistently. As your roster improves your expected finishing position will too

Also, my alliance approves using Thor as a headbutt sponge. Good job

Unknown · March 2014

I'm new here but this is just a hunch..

I think that in these events, the more cumulative points that are earned, the bigger the prize pool gets for each fight you win. Towards the latter few hours, everyone is tussling for that top spot and a lot of points are being earned and therefore the points prizes are getting bigger.

Solely down to the volume of players in the event during the end it kind of snowballs and this makes it more volatile.

For these big global events, waiting is not really advised in my opinion. however perhaps try and keep 4 or 5 health packs available for the end hour or so.

FYI round 1 wasn't great for UK time, but round 2 certainly ends at a good time. Much like myself you're in for a chance for a top spot this time round

Unknown · March 2014

Quick run down of rubberbanding, refreshes and stacks:

Rubberbanding is a mechanism that allows people that are behind to pick up more points compared to how much they would get if they were not as behind. For instance when the event started the first node was worth a certain amount of points (I don't remember how much). Right now for those of us that have not started it is worth 484. If we wait a while longer it will eventually max out at 500. In any case the later you start the higher that first node will be (up to a max).

So, what is rubberbanding based on? Well since I haven't started the event yet that means I'm not in a bracket so I can deduce that the rubberbanding for this particular event is based on the global points leader (the person with the highest number of points, regardless of what bracket they are in).

How do I know this/what were the other possibilities? If there were sub brackets then we could have the rubberbanding be based on the global main bracket leader, your main bracket leader, or the global sub bracket leader. There are no sub brackets for this event so that eliminates the last possibility. And it can't be based on my main bracket leader because I'm seeing the points rise (rubberband) despite the fact that I'm not in a bracket yet.

Well if there is this magical thing called rubberbanding why should I play early? Well there is a max to the rubberbanding so that (usually) limits the amount of progression rewards you could get. And everyone else that is in the same position as you (in total event points) will see the same amount of rubberbanding. But most important is the fact that rubberbanding will not allow you to beat someone who played more than you. For instance if I play optimally (see below) I'll be getting lower point values for each mission but your higher rubberbandeded values will not allow you to catch me if you happened to miss a few missions or didn't play optimally (again see below).

Playing optimally: Stacks and Refreshes:

The point values for each missions are represented by stacks. So let's say that first mission has a stack of 500-400-300-200-100-1-1-1-1... (I don't remember what/how many it is for this event). This means the first time I play it I will get 500 for a win, the second time I'll get 400 and so on.

The stacks refresh (for this event every 12 hours) so lets say I grind that same stack down to 1 point at 11am (so I've completed that mission 5 (or whatever it is for this event) times, the last of which ended at 11am). Then at 11pm the stack will be refreshed. Now to be clear all that means is you will again be able to play it 5 (or whatever) times before the value drops back down to 1. I can't say if it will refresh back to 500 because that number depends on rubberbanding which as explained above depends on your total event points and the total event points that the global point leader has (which is next to impossible to know). But that doesn't matter, because after those 12 hours have passed your points are what they are, if you wait a while to let the global leader get some more points that value will go up because of rubberbanding.

So when should I play the stack after it has refreshed? This is where we really start talking about playing optimally. First an oversimplification: pretend we know for sure that grinding the stacks for every mission in the current event will take us two hours (a pipe dream for this event but we're pretending). That means two hours before the event ends I want all my stacks to be refreshed. So I want to be ending my second to last refresh/play through 14 hours before the event ends (12 hours to refresh + 2 hours to grind them all again). So that means I want to start my second to last play through 16 hours before the event ends (2 hours to grind, 12 hour refresh, 2 hours to grind). Now back track in this way to when the event starts and you can figure out all the optimal times to play.

Why did you call that an oversimplification? Simply because finishing the last mission doesn't have to be done according to that exact schedule. If it takes you 10 minutes to grind the last mission down then you really only need to be starting it (for your second to last refresh) 12 hours and 20 minutes before the event ends (10 minutes to grind, 12 hour refresh, 10 minutes to grind). Figure out a schedule with all of that in mind for each mission and you can truly play optimally.

But why? Rubberbanding! You need to wait as long as possible to start grinding any particular refresh in order to get the most points from it and the you want the most refreshes possible otherwise you're not playing optimally.

Disclaimer: I'm not an expert, this is just how I've seen things play out. I'm sure I overlooked certain possibilities that will affect you in certain situations and for that I apologize. I'm also sorry for now knowing all the precise numbers/values for this particular event. But if you do know those you should be able to use this for any event that is set up how the pve events are nowadays. I'm sure if anyone takes the time to read this they will be more than willing to point out the mistakes.

The most important things to figure out are: the refresh rate (someone is always willing to start early/take one for the team and report their results on the forum), what rubberbanding is based on (global event leader, main bracket leader, global sub bracket leader), how long it will take you to do a full play through of the missions, how much of that full play through you want to do (a full play through is every node ground down to 1 point).

So why doesn't everyone play optimally all the time? Life, time constraints: significant others, work, sleep, sanity, also health packs are a concern, does your optimal play through time take in to account outside means of healing, an irrational fear of countdown tiles, a rational fear of headbutts, etc...

Pwuz_ · March 2014

I'd only like to add that it seems some events (like this one) actually seem to have all mission points decrease when you finish each mission. i.e. I start playing Mission 12 which is worth 500 points, but avoid mission 11 since it's only worth 480 points, after grinding down Mission 12 to 1 point, Mission 11 has also dropped down to 100 or so. Though this may be due to proximity to global leader rather than just arbitrarily dropping after every mission, I don't know, just pointing out an observation.

Thanks for the proper explanation beezer37_84.

Unknown · March 2014

Pwuz_ wrote:

Though this may be due to proximity to global leader rather than just arbitrarily dropping after every mission

Yep, it's all due to rubberbanding.

And that's why it's important to figure out what the rubberbanding is based on. For instance in the last Hulk event it didn't matter which of the sub events you worked on first because rubberbanding in the subs was based on the global sub bracket leaders for each particular sub.

Unknown · March 2014

beezer37_84 wrote:

Quick run down of rubberbanding, refreshes and stacks

Hey Beezer37_84, first off a big thank you to your guide on this. Seriously dude you should right professionally that kinda came off as convoluted but was incredibly easy to understand with your explanation. Second no need to apologise on anything, this is a very good guide that lets me know it does take effort to make it to the top ranks especially on these PVE events, and also that my previous theory on how I could hit the top spots was wrong. Probably also explains why it didn't work out for me last time seeing as I didn't really play too much of the event until the day it was expiring.

Taking all of your advice on board for future events. I also understand that if there are more higher tiered players in my bracket, it makes the top rank so much more difficult, but it's just so much more comforting to know there is logic (maybe flawed in some ways) behind the system.

Lastly I have to take you for your word considering you're part of the S.H.I.E.L.D alliance (side note, your abbreviation explanation for SHIELD made me laugh), you guys wouldn't be the top players because you didn't have a good idea on the systems.

Once again, big thank you for the explanation into these systems, much appreciated.

Unknown · March 2014

Thanks for the kind words!

Supermans Doctor wrote:

I also understand that if there are more higher tiered players in my bracket, it makes the top rank so much more difficult

There is another mechanism we should talk about: scaling. This affects the levels of enemies you see but it's still fairly new and is being tweaked (as with most other things). It's based on a lot of stuff in general: how you've been performing in pve events (referred to as your pve mmr), how every other player is doing on a specific mission, how you've been doing on that mission, same for the current event, and other factors (floor, ceiling, etc...) that are set by the devs. In a perfect world this would match you up against enemies in a range of difficulty decided upon by the devs for each individual mission.

So far in practice (and in general) what we've seen is beginning players get fairly easy set ups and more advanced players get high level enemies which for the most part their more advanced teams can deal with but it becomes a big slog. So far (anecdotally) mid-level players are the most adversely affected because they get a good amount of scaling up in enemy levels but their teams aren't as advanced.

What is making this particular event even more difficult for some is that the rosters are restricted so they might not have their usual bag of tricks available to them.

But all that aside I wanted to mention this because if scaling was perfect (I'm not saying the use of such a system is perfect) the person that would win would be the one that plays optimally as previously discussed. So in practice (under a perfect scaling system) if you saw Nemek or Jozier or another heavy hitter in your bracket you wouldn't be worried about them getting first place because of their team, but rather their dedication (which in turn has earned them those scary teams that they have).

Supermans Doctor wrote:

your abbreviation explanation for SHIELD made me laugh

Full credit for this goes to Clint (clintman in game)

Unknown · March 2014

beezer37_84 wrote:

Playing optimally: Stacks and Refreshes

I'll have you know, I came back just to rep this. Thanks again.

Nemek · March 2014

Oh dang, I was just going to refer you to the FAQ I made from the Hulk Event, which I thought did a pretty decent job, but Beezer has an incredible post.

But uhh...here's the link, anyway:

http://www.d3pforums.com/viewtopic.php?f=7&t=1737

Unknown · March 2014

Nemek wrote:

Oh dang, I was just going to refer you to the FAQ I made from the Hulk Event, which I thought did a pretty decent job, but Beezer has an incredible post.

But uhh...here's the link, anyway:

http://www.d3pforums.com/viewtopic.php?f=7&t=1737

One free interweb for effort!

agent86ix · March 2014

I'm not nearly as hardcore as some are on this forum. My goals are much more modest, but I'll share my strategy for those who are "lazy" like I am.

To start with, these are my "goals-slash-accomplishments" while playing MPQ:

I don't pay money or sink lots of hours into the game. If I play for an hour a day, that's kind of a lot. I generally try to get at least one cover for the "new hotness" and hang onto it for a few weeks while the wave of "events where that guy/gal matters" washes over the game. After a while, I'll generally dump characters if I need slots or if they're not particularly rare. Progression-wise, I'm transitioning from 2* to 3* and so far it's been going OK. I've got a lot of 1/2/3 cover 3* heroes, and a few stand-outs (Patch, Punisher).

In the last Hulk event, I topped out the progression rewards (got the punisher covers! yay!) and got one "good" token from each sub event. I finished high enough to get a 3* Thor cover, which I'm using in the current Heroic Juggernaut event. I've been in the top 100 in my bracket (which means I get another 3* Thor cover). I also maxed out all the current event's progression rewards. This is kind of the "sweet spot" I've generally shot for - there's some spot on the graph of time invested vs. rewards gained that I try to find and live in. Mostly I'm going for covers. The recent shift in balance more towards ISO than covers has been bad for me, unfortunately.

My strategy is generally:

- Play through the missions once. (Somewhat optional, if you have the "featured required character")
- Grind the highest value missions until they are below some threshold (usually this is 500 points for me, unless it's coming close to the end and I'm close enough to hit some progression reward I'm missing)
- Let time pass, and wait for the scores to recover
- Lather, rinse, repeat until the end of the event is near
- Grind as much as I can bear until I meet the leaderboard/progression goal, or until I decide it's unattainable

This relies on people being willing to grind, grind, grind for the leaderboards, and for the most part it seems there are some people passionate enough to do so in every event. I generally land comfortably in the leaderboard reward group I plan to hit.

By not grinding every node down to the nub every time it refreshes, I avoid having really hard fights near the end. In the current round of the Heroic Juggernaut event, I'm actually only playing the "3* Thor required" nodes, and they're all still listed as "Trivial." I haven't even unlocked 80% of the other nodes, I think I played the first two. Right now I think I'm in the 50's as far as the leaderboard is concerned.

It's not a strategy that is going to get you to the top, and I'll admit it is very lazy when compared to the "optimal" way of playing described in this thread. However, given the amount of time I have to play, I find it to be "optimal" for my purposes.

Cousin Simpson · March 2014

agent86ix wrote:

By not grinding every node down to the nub every time it refreshes, I avoid having really hard fights near the end. In the current round of the Heroic Juggernaut event, I'm actually only playing the "3* Thor required" nodes, and they're all still listed as "Trivial." I haven't even unlocked 80% of the other nodes, I think I played the first two. Right now I think I'm in the 50's as far as the leaderboard is concerned.

I'm going to highlight this cos it might get buried in the rest of the post. Very interesting! However, I have a question for you: did you use this strategy to get the Punisher covers from the Hulk PVE? Does it work to advancing quickly through progression rewards?

agent86ix · March 2014

Cousin Simpson wrote:

I'm going to highlight this cos it might get buried in the rest of the post. Very interesting! However, I have a question for you: did you use this strategy to get the Punisher covers from the Hulk PVE? Does it work to advancing quickly through progression rewards?

Yes, I used this to get the Punisher covers in this most recent Hulk event. I got the last one I think a day or two before the event ended. 3/4ths of the way through I started to get a bit paranoid I might not make it - I think there were a few I played down till they were worth 400 points or so, but not generally below that point.

On that one too, late in the event I was mostly playing the Black Panther and Punisher nodes. (I had I think 1, maybe 2 Black Panther covers at the time.) The rest of my team was generally Cap'n A and OBW, both ended at 85 although OBW was in her 70's when I started the event. I was generally crushing those nodes. They got harder as the sub-events wore on, and by the end it was getting a bit tricky.

I think I read an IceIX post that indicated that the scaling ramps up if you're winning with no/light damage. In the current event, I'm taking moderate damage (occasionally losing a character, although we're talking Modern Storm here) and using health packs between fights. I think that might be why I'm not seeing the scaling as much this time around.

Cousin Simpson · March 2014

agent86ix wrote:

Cousin Simpson wrote:

I'm going to highlight this cos it might get buried in the rest of the post. Very interesting! However, I have a question for you: did you use this strategy to get the Punisher covers from the Hulk PVE? Does it work to advancing quickly through progression rewards?

Yes, I used this to get the Punisher covers in this most recent Hulk event. I got the last one I think a day or two before the event ended. 3/4ths of the way through I started to get a bit paranoid I might not make it - I think there were a few I played down till they were worth 400 points or so, but not generally below that point.

On that one too, late in the event I was mostly playing the Black Panther and Punisher nodes. (I had I think 1, maybe 2 Black Panther covers at the time.) The rest of my team was generally Cap'n A and OBW, both ended at 85 although OBW was in her 70's when I started the event. I was generally crushing those nodes. They got harder as the sub-events wore on, and by the end it was getting a bit tricky.

I think I read an IceIX post that indicated that the scaling ramps up if you're winning with no/light damage. In the current event, I'm taking moderate damage (occasionally losing a character, although we're talking Modern Storm here) and using health packs between fights. I think that might be why I'm not seeing the scaling as much this time around.

Huh! Then my understanding of the point mechanics is way off! You play less but somehow get more points, in other words? Is it because PVE MMR sees that you're falling behind in progressions and thus dangles huge point totals for you to catch up?

agent86ix · March 2014

Cousin Simpson wrote:

Huh! Then my understanding of the point mechanics is way off! You play less but somehow get more points, in other words? Is it because PVE MMR sees that you're falling behind in progressions and thus dangles huge point totals for you to catch up?

Well, I don't know that I was earning more points overall than most - in the Hulk event, I was around rank 3,500. Enough to get the 1 3*Thor cover, and spitting distance from the next reward bracket, but I think I started to fall off towards the end. my only goal was above rank 6,666 so I could get one 3* Thor. I'm sure people who were grinding the heck out of the event managed to do better.

I also had both Punisher and Black Panther going in, and the "flavor of the month" hero is always the one that has the high-point-total "required" nodes. Thus, my time investment mostly revolved around BP, and Punisher to a lesser extent. This may have put me ahead of people who were grinding willy-nilly but probably not in front of folks who were doing what is advocated in this thread.

If I left those Pun/BP nodes alone, they'd generally be worth ~1.5-2k after a few hours. I could usually play them repeatedly at least twice a day for the major points. There's certainly a cap to the amount they'd be worth, but I couldn't tell you exactly what the formula was.

mouser · March 2014

Cousin Simpson wrote:

Huh! Then my understanding of the point mechanics is way off! You play less but somehow get more points, in other words? Is it because PVE MMR sees that you're falling behind in progressions and thus dangles huge point totals for you to catch up?

The point rubber banding mechanic appears to work something like the following. Note: The devs are very possibly tweaking this from event to event, so even if it's accurate now it might not be so next event.

- The game determines a point score for the "leader".
-- We don't know for certain whether the leader is the person with the highest point score, or perhaps some averaging of the top X number of players. But either way it works out similarly.
-- In the case where rubber banding is implemented in sub-events that "leader" could be based on the global sub-event leader, local sub-event leader, or main bracket total score leader. The latest Hulk event appears to have used global sub-event leaders for the sub-events and the total score leader for rubber banding in the main bracket missions. Some past events have been implemented differently.

- Every mission has a base point value.
- The game determines how far behind the "leader" you are and proportionally multiplies mission scores based on this.
- The maximum rubber band bonus is 900% -- making the maximum rubber banded mission score 10x the base value (1x base + 9x bonus)
- If you are over a certain amount behind the leader you will receive the maximum +900% bonus.
- If you are still behind the leader, but by less then the threshold you will receive a bonus proportional to the difference.
- If you are the leader then you'll only get the mission base value.
- For example, in the last Hulk event the main bracket bonus threshold appears to have been 6k-7k. So if the max threshold was 6k and you were 4k behind the leader then your mission bonus would be 4k/6k * 900% = 600%. Making your missions have a 7x total value.
- As you complete missions you'll normally be gaining ground on the leader, meaning your overall rubber band bonus multiplier will usually be decreasing during a play session. The bonus will usually increase while you're not playing, assuming the leader is gaining points.

Example:
Using our previous example of being 4k behind the leader with a 6k maximum bonus gap, making your current bonus +600%. Let's say you have two available missions with bases of 100 and 150. With the +600% the vales you're presented with they're worth 700 and 1050 respectively. Let's also assume that our missions are fully refreshed with the most commonly used 100%, 80%, 60%, 40%, 20% repeat values.
- Let's say you do the 1050 first. You will now be 2950 points behind the leader and your bonus will be 442.5% (2950/6000). Your available missions will now be 100(+442.5%) = 542 and 150(+442.50%)*80% = 651 points.
- Let's say you do the 700 first. You will now be 3300 points behind the leader and your bonus will be 495% (3300/6000). Your available missions will now be 100(+495%)*80% = 476 and 150(+495%) = 892 points (rounded down here, but might round up instead).
- And for those keeping track, you might notice that you end up with the same # of points regardless of the order you do those two missions in. 1050+542 = 1592 in the first example, and 700+892 = 1592 in the second example.

Nemek · March 2014

I'm going to plug a different thread of mine that says very similar things to mouser's post:

http://www.d3pforums.com/viewtopic.php?f=7&t=2945

(mouser's post was excellent, btw)

agent86ix · March 2014

mouser wrote:

- And for those keeping track, you might notice that you end up with the same # of points regardless of the order you do those two missions in. 1050+542 = 1592 in the first example, and 700+892 = 1592 in the second example.

I'll point out that using my "lazy strategy" that all else being equal (or close to equal), you'd want to play the hardest mission above the threshold first. My strategy optimizes for scaling and time. Thus, if you play the easy mission and then the hard mission, there's a chance that your scaling will change after playing the easy mission, which would make the hard mission harder. From a point standpoint, it wouldn't matter, but if you're right on the ragged edge of being able to clear a mission, having it jump up and get harder on you before you get to it is a bummer

In the current event, I'll generally play the "hard mode" missions prior to the "not-hard-mode" ones. I'm not 100% sure the scaling is shared between these two events, but from my experience it seems to be so.

Unknown · March 2014

agent86ix wrote:

mouser wrote:

- And for those keeping track, you might notice that you end up with the same # of points regardless of the order you do those two missions in. 1050+542 = 1592 in the first example, and 700+892 = 1592 in the second example.

I'll point out that using my "lazy strategy" that all else being equal (or close to equal), you'd want to play the hardest mission above the threshold first. My strategy optimizes for scaling and time. Thus, if you play the easy mission and then the hard mission, there's a chance that your scaling will change after playing the easy mission, which would make the hard mission harder. From a point standpoint, it wouldn't matter, but if you're right on the ragged edge of being able to clear a mission, having it jump up and get harder on you before you get to it is a bummer

In the current event, I'll generally play the "hard mode" missions prior to the "not-hard-mode" ones. I'm not 100% sure the scaling is shared between these two events, but from my experience it seems to be so.

I thought I was the only one doing this. It's amazing that you only play an hour a day and still get good rewards. I play for over 3 hours a day. I managed to get 2 Lazy Thor covers during the Hulk event using the "lazy strategy" and I only played 7/9 days of the event. Like agent86ix, I also play the hardest mission first, although what I do is I don't grind nodes one at a time. For the simulator, I'm not sure if it actually procures more points but I only play the nodes with essential characters until the points are lower than the regular missions... starting with the left and then right (either way works for me since they both have the same amount of points) then the middle... lather, rinse, and repeat. I also start the two subs 4 hours apart so my health packs would regenerate and I could focus on each sub completely. Although I didn't realize that playing the easy mode first would make the hard mode harder... Duly noted. Will reverse order on the next set.

The secret art of rubberbanding

Comments

Categories