Ideal PVE Clearing Order.........by Mathematical Analysis

2»

Comments

  • bbigler
    bbigler Posts: 2,115 Chairperson of the Boards
    The clearing times I used are very close to my actual play.  Of course, each event will have variations, but I just needed an average time.  The 3rd node for me takes 3 times longer than the 2nd because I don't use 3*Thanos with 6K+ health enemies.  The 6th node is easier for me than the 4* node because I'm not limited/handicapped by the required character.

    Just for interest: I'm currently ranked #11 in SCL8 ISO-8 Brotherhood and #10 only has 1 more point than me, and #9 only has 2 more points.  So, squeaking out a few more points with ideal clearing order is definitely worth it. 

    I fully believe my methods are sound because I go up in rank at the end of the sub-event in comparison to my rank at the beginning.  There are people that clear faster than me, but I believe I'm clearing in a better order than them. 
  • Yepyep
    Yepyep Posts: 952 Critical Contributor
    Yepyep said:
    bbigler said:

    It's always better to finish too early than to finish too late and completely miss some 7th clears.  
    Ok, this confuses me. 7 clears? There are only 6 available. What are you talking about...?
    There are only Rewards for the first 6 Clears, but you can keep playing green checked Nodes for more Points. You can get one more Clear (for a total of 7) worth a significant number of Points from each Node in a one-day Sub, and two more (for a total of 8) in a two-day Sub.
    Thank you! I made good use of that info in the current event. I never knew this...
  • TPF Alexis
    TPF Alexis Posts: 3,826 Chairperson of the Boards
    Yepyep said:
    Yepyep said:
    bbigler said:

    It's always better to finish too early than to finish too late and completely miss some 7th clears.  
    Ok, this confuses me. 7 clears? There are only 6 available. What are you talking about...?
    There are only Rewards for the first 6 Clears, but you can keep playing green checked Nodes for more Points. You can get one more Clear (for a total of 7) worth a significant number of Points from each Node in a one-day Sub, and two more (for a total of 8) in a two-day Sub.
    Thank you! I made good use of that info in the current event. I never knew this...
    You're welcome! I didn't find it out myself until I started hanging out on the forum and someone mentioned it.
  • robertbah
    robertbah Posts: 124 Tile Toppler
    bbigler said:
    If you want to re-create my tests, here are the numbers that I used:

    Refresh Rates (points/minute)
    1 - 0.06159
    2 - 0.08333
    3 - 0.10145
    4 - 0.11957
    2* - 0.13043
    5 - 0.14130
    3* - 0.14130
    4* - 0.15580
    6 - 0.16304

    Starting (full) points per node:
    1 - 270
    2 - 355
    3 - 439
    4 - 522
    2* - 556
    5 - 609
    3* - 614
    4* - 678
    6 - 695

    Time to clear a node once (minutes):
    1 - 0.5
    2 - 0.5
    3 - 1.5
    4 - 2.0
    2* - 2.5
    5 - 3.0
    3* - 3.5
    4* - 3.5
    6 - 3.0
    Note: node clearing times are shorter during the first 4 clears.
    Ideally, the order of initial clears should be based on the rate = points/time. The node with highest rate should be cleared first. In your setup, the rates are
    1 - 540
    2 - 710
    3 - 292
    4 - 261
    2* - 222
    5 - 203
    3* - 175
    4* - 193
    6 - 232

    Therefore, the best way is in fact 2,1,3, ... and so on.
    Of course, for 1-6, you need to do one fight to open the nodes, that will decrease the rate for later nodes 4-6 even further. This principle could be used for 2* 3* 4* because they are available right at the beginning. In this case 2* 4* 3* should be the best order. In practice, you do not know how fast you can clear, so you will not be able to have estimate of rate before you actually play the nodes. My conclusion is that OP's analysis is interesting. For me, I will just pick the order I am comfortable with. I prefer to use the simple method and decrease the number of team changes. It takes time to change time, too.

  • Pongie
    Pongie Posts: 1,410 Chairperson of the Boards
    I don't think the refresh rate matters that much when it comes to the order of the first 4 clears. You don't get fractions of points when clearing. The difference in rates are so small that it takes several minutes before you get 1 point.
  • Pongie
    Pongie Posts: 1,410 Chairperson of the Boards
    bbigler said:
    Just for interest: I'm currently ranked #11 in SCL8 ISO-8 Brotherhood and #10 only has 1 more point than me, and #9 only has 2 more points.  So, squeaking out a few more points with ideal clearing order is definitely worth it. 

    I fully believe my methods are sound because I go up in rank at the end of the sub-event in comparison to my rank at the beginning.  There are people that clear faster than me, but I believe I'm clearing in a better order than them. 
    I was #11 in SCL8 for ISO-8 Brotherhood too (about 2 subs ago), and now I have managed to somehow climbed to #5. Though I don't think it was entirely using this new grind order. Every sub I have been finishing top 10 but never higher than top 5. I believe, some of those that were in top 10 decided enough was enough and stopped playing optimally. 7 days is taxing, especially when a new release event is coming up next. Kind of great for me, since I need top 5 placement reward to finish iron fist (no ISO to champion him though).
  • bbigler
    bbigler Posts: 2,115 Chairperson of the Boards
    edited June 2017
    PVE events show no mercy!  Last night, I was 2 points behind #9 before starting my grind.  I started my grind with 1 hr and 2 min remaining, which seemed possible given my clearing times from the last few nights.  Well, my matches were tougher than expected and I ended up completely missing some 7th clears!  No big deal, you say?  Well, when it was all done, I was 500 points behind the top 10, which means I have no chance at getting Hobofist covers despite playing perfectly for the last 3 days. 

    That's the frustrating part of these 7 day events: you have to play perfectly for 7 days straight to get top 10 placement.  So, it feels like all my hard work was lost because I mis-judged my clearing times.  My only chance now is to hope that a few people higher in rank make the same mistake that I did.  This reinforces my statement that it's better to clear too early than too late.
  • nigelregal
    nigelregal Posts: 184 Tile Toppler
    I have been playing almost perfectly for this event (I did a test of a 9 hour pre grind though). I'm currently in 1st by 4000 pts I think. I am in CL7 though. 

    If I was in CL8 I would be top 20 I think for each sub but in CL7 i'm 1st every sub. So I am getting an extra 5 CP per sub for sacrificing the 8 or so CP from dropping to CL7
  • Jesus Jones
    Jesus Jones Posts: 182 Tile Toppler
    I misjudged 2 nights in a row and didn't complete all levels, totally killed myself!!
  • Pongie
    Pongie Posts: 1,410 Chairperson of the Boards
    Im back in #6 now (#4 didn't bother clearing at the start of the sub). #5 gap has widen from ~60 to ~100 so I'm not making any gains. My only hope is #4 would be slow with the grind too
  • TheLadder77
    TheLadder77 Posts: 40 Just Dropped In
    edited June 2017
    bbigler said:
    PVE events show no mercy!  Last night, I was 2 points behind #9 before starting my grind.  I started my grind with 1 hr and 2 min remaining, which seemed possible given my clearing times from the last few nights.  Well, my matches were tougher than expected and I ended up completely missing some 7th clears!  No big deal, you say?  Well, when it was all done, I was 500 points behind the top 10, which means I have no chance at getting Hobofist covers despite playing perfectly for the last 3 days. 

    That's the frustrating part of these 7 day events: you have to play perfectly for 7 days straight to get top 10 placement.  So, it feels like all my hard work was lost because I mis-judged my clearing times.  My only chance now is to hope that a few people higher in rank make the same mistake that I did.  This reinforces my statement that it's better to clear too early than too late.

    I've been using that 3 wave clear method for final clears for several months now, nice to see somebody has taken the time to lay out the maths. I've never been a scientific player, it just struck me as the sensible way to do things.

    On your story there, you're right - it is much better to leave too much time for clears than not enough. That is a much harder part of the game to get right, not sure if there is a scientific way of dealing with it. That one bad match that stalls you is hard to account for.
  • Shin Tenma
    Shin Tenma Posts: 6 Just Dropped In
    What I have found about judging the clear times is that you should give yourself and extra 2 or 3 clears worth of the 6th mission time for the final 3 (or 4) clears on the sub. If things go poorly you have time to make sure you get at least your full 7 clears in. If things go as expected then you can get in a couple extra clears of either the 6 or 4* missions in (whichever goes quicker). If things go better than expected then you have even more time to get a few additional clears in before the end of the sub.

    For example: I expect that tonight's final 3 clears of the Iso-8 Lab: Canada sub to take me 45 minutes. I know that mission 6 takes me about 3.5 minutes to clear so I will make sure to add about 10-12 minutes to my clear time. So I'll start my final clears 55-57 minutes before the end of the sub.
  • Big10Inch
    Big10Inch Posts: 23 Just Dropped In
    edited February 2018
    Does anyone know if the node points run the same still, as I think they've changed and the 2* is worth more than 6?

    Also, for anyone looking at this there seems to be quite a few errors in data and calculations above. If the first node is work 270, then each clear is worth 90, with a regeneration rate of 0.0625. This runs through all the above, with, I think, a few copy and paste errors.

    I'm doing some data modelling on this now, as I like to geek out :-)

    EDIT

    OK Modelling complete. I've worked through 20 permutations between the first 4 clears, and the final 3 clears, and what it shows is the best way to start is to run through low to high - not high to low as per the original post.

    Since this thread started the points have changed and the scores are more linear, so points from low to high are now simply: 1,2,3,4,5,6,2*,3*,4*,5*. I'll check this again on Prodigal, but it has been on the last few events.  
     
    So low to high means run your first four clears as 1,1,1,1,2,2,2,2,3,3,3,3,4.....5*.

    While this is counter-intuitive, as you're leaving the highest points until last, in fact what you're doing is getting more nodes to the countdown/refresh point quicker. Plus there's no time wasted as you can do a straight run through. You could try and work out the best points versus time as has been talked about, however I wanted to look at easy to implement models.

    For your final 3 runs, on my modelling so far, it's a tie for two models, both have node clears all together, ie 1,1,1... not in waves, 1,2,3... . Initially the best option was to run high to low (5* first), however I re-modeled based on the points, and my times, for the last sub-event in Wakanda, and the results gave the same figure for low to high. I'd still recommend the first option though, as if you finish early you can go back and get another 20 points or so on the 5* node.

    All of this means a whopping 123 point difference, 150ish if you get the extra 20, so it's not worth worrying about unless you're chasing a place. The main variable remains how quickly you can clear them all.   
  • Yepyep
    Yepyep Posts: 952 Critical Contributor
    Big10Inch said:
    ... I think they've changed ...
      
    It's always worth going back and checking old assumptions! Thanks for this. Nothing says they can't, or won't, change anything at all in the game. It is, after all, their property. (For example, it is my pretty firm conviction that they have made pretty substantial "improvements" to the matching algorithms over the last 3-4 months. I don't know how this could be tracked, so it is consequently anecdotal, but nevertheless it is what I believe.)
  • Pongie
    Pongie Posts: 1,410 Chairperson of the Boards
    edited March 2018
    Big10Inch said:
    If the first node is work 270, then each clear is worth 90, with a regeneration rate of 0.0625. This runs through all the above, with, I think, a few copy and paste errors.
    If you’re referring to the 5th, 6th and 7th hit, then no these aren’t 1/3 of original value. 5th is 2/3 of original, 6th is 2/3 of 5th, and 7th is 2/3 of 6th. Each clear reduce the value by a third. Hence the longer you let it refresh, the more value 5th-7th clears becomes. 

    Edit: if the node values change, all that’s done is alter the order. The same strategy still applies. Not that any of this matters. Tappers will beat you anyway. 
  • Big10Inch
    Big10Inch Posts: 23 Just Dropped In
    Pongie said:
    Big10Inch said:
    If the first node is work 270, then each clear is worth 90, with a regeneration rate of 0.0625. This runs through all the above, with, I think, a few copy and paste errors.
    If you’re referring to the 5th, 6th and 7th hit, then no these aren’t 1/3 of original value. 5th is 2/3 of original, 6th is 2/3 of 5th, and 7th is 2/3 of 6th. Each clear reduce the value by a third. Hence the longer you let it refresh, the more value 5th-7th clears becomes. 

    Edit: if the node values change, all that’s done is alter the order. The same strategy still applies. Not that any of this matters. Tappers will beat you anyway. 
    That's not my understanding, and not what I've seen. For example on the Prodigal Sun event, when we had the required character node running for the duration, each day it returned to max, and dropped by 1/3 on the hit. If the node dropped by a third of current value it would never reach 0 (or fixed minimum).

    The 5th attack should be near full points, not 2/3, and the points drop by a third each time. So simplifying it: Node worth 900pts, each attack drops by 300, down to 0, so
    5th - 900
    6th - 600
    7th - 300
    Then 0 or whatever the remaining value is (and I see they're trying a no tapping approach with this event).

    On your version the points would be:
    5th - 600
    6th - 400
    7th - 266
    8th - 177
    9th - 118 etc.
       
    If someone has the time, and roster, to tap out enough points to win, then lucky them, but for a lot of us RL gets in the way, so I'd rather maximize the points from my playing time.
  • robertbah
    robertbah Posts: 124 Tile Toppler
    Node value = full_value * (1- timer/72hr) 
    each clear will add 24hr on the current timer. 
    Node value will be maxed when timer is less than 4m19s. (This works 90% of time. Except one event)