data I collected on rubberbanding

Unknown

So in another thread I made the suggestion that the sub bracket leader doesn't change often and it's a valid question so I collected some data to see whether the overall sub bracket leader does indeed change or not.

With 5H 30M left on the tournament, I have 94 on the simulation 1 (base 50) on the normal bracket and 200 on simulation 1 (base 100) on the hard bracket.

With 4H 45M left on the tournament, I have 106 on the simulation 1 normal and 203 on simulation 1 hard.

So waiting for 45 minutes you'd get 12 more points on the smallest mission on normal, and 3 points on the smallest mission on hard. Except in that time, you can easily get more points than that if you actually played the game.

Note that the movement on the hard bracket is much smaller (200 to 203) than the normal bracket (94 to 106), which is hardly surprising.

In fact given the time it takes to do the 230X3 missions, it'd make very little sense to push for more when you're about to tie the overall sub bracket leader, since that's obviously enough to hold #1 in your sub bracket and there's no reason for you to work extra hard on those 230X3s. For the easier missions, there might be an incentive to wait a bit, but even then, the point you get for waiting does not necessarily compare to the points you can get by just playing. Yes in theory it doesn't matter if you can finish all the missions, but I don't know of anyone who's done that, and such a person would never need to worry about his placement in the first place.

Eddiemon

The main issue I would have with your analysis is your sample timing. With 4 to 5 hours to go most people who know what they are doing and intend to compete are siting till 2 hours out or less to hit it hard. So you're selecting data in what is effectively the eye of the storm.

During this particular window the only reason to push is if you have to work/sleep and are willing to settle for a score in the 50s if you are lucky.

Unknown

Yep, that looks like a dead time period where everyone is waiting and waiting.

And the compare doe not seem plausible to start with the diff on the fisrt game may be small points but if you engage the multiplier starts dropping, and there you start losing serious bonus and use up the good base points.

With RB at puny 2x it takes just few battles to gain to the top -- but the next guy will overtake you just as easily 30 minutes ahead. At least as I understand this #%#$%# mechanics.

Unknown

My hard bracket number for the first mission has been pretty much been changing by at most 10 points even up to now. The normal bracket numbers went up by about 50, which is basically 10 times the net rubberband effect when you consider the normal bracket base points are half of the hard bracket.

A guy who is currently the overall sub bracket leader is generally comfortably ahead in his own sub bracket. He'd have no reason to further expand on that lead, especially when his enemies are all 230X3. You see more volatility on the normal bracket, because the enemies there are lower so someone who is the overall sub bracket leader may indeed try to increase his lead, but only because the enemies are easy. I've been writing down my simulation 1 numbers as I clear alone and it's virtually unchanged in hard, while there's certainly a lot of changes in the normal bracket.

The argument for starting late seems to be 'he who plays last wins', which is sort of true in some extent, but guess who is the guy who can afford to play last? Whoever is currently #1 in your sub bracket. Unless the #1 for some reason continued to plug away while he's #1, he's the guy who gets to play later than any of his pursuers.

Unknown

I have been trying to figure this out all day to no success. I waited like 10 hours for a magical reset/respawn that never happend...

Unknown

klingsor wrote:

I have been trying to figure this out all day to no success. I waited like 10 hours for a magical reset/respawn that never happend...

Reset is 12 hours.

Unknown

Phantron wrote:

klingsor wrote:

I have been trying to figure this out all day to no success. I waited like 10 hours for a magical reset/respawn that never happend...

Reset is 12 hours.

I know, but it's not discrete 12 hour jumps, it is gradual, so at 10 hrs it should be pretty damn close no?

morgh

Actually - no. Take a node of 1000 base points - every time you deplete it, it goes down by 200 (so you can get 1000+800+600+400+200).
Now let's consider you were tired and only depleted 3 'stacks' -
you got 1000 at 24:00
800 at 00.05
600 at 00.10
so it is now worth 400

You come back after 11 hours 55 minutes (11.55) - and it's still worth only 400... now it goes back to 600 at 12:00 (noon), 800 at 12:05 and 1000 at 12:10

Of course above explanation does NOT include rubberbanding which would make the point totals much higher

Unknown

klingsor wrote:

Phantron wrote:

klingsor wrote:

I have been trying to figure this out all day to no success. I waited like 10 hours for a magical reset/respawn that never happend...

Reset is 12 hours.

I know, but it's not discrete 12 hour jumps, it is gradual, so at 10 hrs it should be pretty damn close no?

It is discrete 12 hour jumps.

mouser

I can tell you with 17 hours to go that the Hard sub event was climbing by a good amount, whether that be from a single global leader or people leapfrogging each other. I was going through a clear at that time while checking my multiplier after each mission and could see the multiplier jump many times.

Unknown

Finished 2nd/2nd this round. After looking at my sub brackets and overall brackets and the guys I'm competing, here's what I concluded. Starting late is advantageous if your PvE scaling is low. The rubberbanding is fairly irrelevent. If there is no PvE scaling and everyone fights level 230X3 enemies, you can basically assume each battle takes about 10 minutes (they take a while no matter who you got). Although no one knows the formula of the rubberbanding for sure, there exists a way to convert a lead into an equivalent base points. That is, maybe for this event 5000 points is the same as having 1000 base mission points lead, and you can convert that to a minimum number of missions, and if it takes at least 10 missions to do that then you know nobody can catch you from 5000 behind with less than 100 minutes left. In these scenarios you basically want to do as many and as often as you can, which is how the pre scaling PvE content worked.

However, because of PvE scaling, the time it takes to defeat an encounter is not always the time it takes to defeat 230X3, especially at the beginning of the event. Therefore if for whatever reason you have a favorable PvE scaling, you want to play as few missions as possible not because of rubberbanding but that you don't want to increase the your scaling. Therefore, a guy who normally has built up say 1000 base mission points of lead cannot hold you off at the end because you're not fighting 230X3s so getting 1000 base mission points takes considerably less than expected. Of course, in such a case you've a rather significant advantage compared to the guy you're attempting to catch, so you probably will win anyway, but you want to play as few games as possible to minimize the increased scaling factor.

Note that on an event like heroic Juggernaut where the scaling seems to be constant, there's no particular advantage for starting late since you're fighting relatively constant level enemies. It only matters on events you know will definitely eventually scale to 230X3s, which is usually the 10 day events.

Now, what if you're someone whose scaling is unfavorable? You pretty much have to just grind as much as you can to drive up the community portion of the difficulty. There's no way you can possibly fend off someone with say a level 15 Psylocke boosted to level 105 fighting level 20-30 enemies while your guys are fighting 230X3 even if you have every character in the game maxed, so you want to build as big a wall for your pursuers so that they don't get a free ride the next event with their favorable PvE scaling. The more your pursuers have to win, the more their own scaling will go up as well, and since your scaling can't possibly go up anymore if you're facing 230X3s, this will help you in the long run.

Unknown

mouser wrote:

I can tell you with 17 hours to go that the Hard sub event was climbing by a good amount, whether that be from a single global leader or people leapfrogging each other. I was going through a clear at that time while checking my multiplier after each mission and could see the multiplier jump many times.

The hard missions were pretty much static from about 5-3 hours to go in terms of increasing point value, and then it went up noticeably with 2h left. That said it didn't affect my standing. I couldn't hold off the guy who had a level 30 as his highest hero fighting really easy enemies and starting late or early wouldn't have mattered there.

Unknown

Phantron wrote:

There's no way you can possibly fend off someone with say a level 15 Psylocke boosted to level 105 fighting level 20-30 enemies while your guys are fighting 230X3 even if you have every character in the game maxed

Saw such a guy. Had a lvl35 IM, 2 other * at 25 a psyloche-red 15 and few under that.
He was in top5 with same points we do.

Unknown

morgh wrote:

Actually - no. Take a node of 1000 base points - every time you deplete it, it goes down by 200 (so you can get 1000+800+600+400+200).
Now let's consider you were tired and only depleted 3 'stacks' -
you got 1000 at 24:00
800 at 00.05
600 at 00.10
so it is now worth 400

You come back after 11 hours 55 minutes (11.55) - and it's still worth only 400... now it goes back to 600 at 12:00 (noon), 800 at 12:05 and 1000 at 12:10

Of course above explanation does NOT include rubberbanding which would make the point totals much higher

What the implication of this is that I needed to stop playing 13 hours before my final clear and not 12 to allow full regen then, no?

Ok, now is it better to clear higher value first, and bring those down, (depth first) or go for highest value of any node, jumping around (breatdh first) or reverse depth first (low to high?)

Unknown

klingsor wrote:

morgh wrote:

Actually - no. Take a node of 1000 base points - every time you deplete it, it goes down by 200 (so you can get 1000+800+600+400+200).
Now let's consider you were tired and only depleted 3 'stacks' -
you got 1000 at 24:00
800 at 00.05
600 at 00.10
so it is now worth 400

You come back after 11 hours 55 minutes (11.55) - and it's still worth only 400... now it goes back to 600 at 12:00 (noon), 800 at 12:05 and 1000 at 12:10

Of course above explanation does NOT include rubberbanding which would make the point totals much higher

What the implication of this is that I needed to stop playing 13 hours before my final clear and not 12 to allow full regen then, no?

Ok, now is it better to clear higher value first, and bring those down, (depth first) or go for highest value of any node, jumping around (breatdh first) or reverse depth first (low to high?)

Breadth first is less work but you have to wait longer. That is, suppose you do A B C A B C A B C, then A will not refresh fully until the refresh period + time spent playing A 3 times, B and C X 2, which can be considerable in it itself. If you did A A A B B B C C C, then A would've refreshed in refresh period + the time it takes to do A 3 times. However doing depth first means you've to know exactly how deep to go on each mission. That is, if you do A A A B B B C C C and found that you still don't have quite enough, unless one more C is enough to stop, you'd have to go back to A or B and that ends up similar to breadth first.

Eddiemon

klingsor wrote:

What the implication of this is that I needed to stop playing 13 hours before my final clear and not 12 to allow full regen then, no?

Ok, now is it better to clear higher value first, and bring those down, (depth first) or go for highest value of any node, jumping around (breatdh first) or reverse depth first (low to high?)

There is a whole Math thread on this stuff.
If your multiplier is 10, then low stuff.
As soon as your multiplier dips below 10, high stuff. (If you could predict exactly when it would drop from 10 doing it as close to that edge as possible is better, but we don't have the technology for that)
Once your multiplier trickles out to 1, whatever is left.

And yes you are best off fully regenerating your high value missions right at the end of the sub.