Author Topic: Pie Factory Experiment  (Read 18078 times)

0 Members and 1 Guest are viewing this topic.

lakeman421

  • Guest
Pie Factory Experiment
« on: August 15, 2015, 11:59:32 am »
Lately I have been experimenting with different pie factory strategies to gain a higher average for that board.  Where I currently stand gaining some points on that board would be crucial.  I also need to prevent losing thousands of points from a bad setup.  This could easily be the difference between a 57k-60k level and a 60k-63k level.  While getting these points I have to take some risks and not lose points to the timer.  I was messing around with certain strategies when I was having issues with the board.  The prime case being a pie spawning out right away, the pie blocking me to the hammer, and fireballs coming down at the same time.  Even if I could grab the hammer in that case chances are i lose points to the timer.  So this is what I started doing.  If I saw the pie come out, I would maybe split a fireball or two just so i could have access to the hammer.  Depending on the scenario, I could even smash the split fireball and send it back so they would all be on the left side of the board.  If the early pie doesn't come out and the board looks clean, than I may look for the free pass.  I used that strategy in both my 1.144 and 1.158 games and averaged about 8700 points per board.

So this is the new strategy I just worked on recently.  Since I forfeit the free pass almost every time now to use the bottom hammer, I decided to just split the fireballs and go for the hammer. This eliminates the possibility of having the occasional free pass with the previous strategy, but it also eliminates the possibility of that first pie getting in the way of the hammer, which might result in wasted time.  Also with this strategy there is the possibility of having a good hammer and clearing the board with around 5800 on the timer.  Other strategies i used may end the board with 5500 or 5600 on the timer after grabbing the bottom hammer.

Once I split the fireballs and go for the hammer this is when I read the board and decide how to play it out.  The best case scenario is seeing three pies coming out right away and I can smash them and go the left before the conveyor switches directions.  While using the hammer I try to set up in case a fireball comes down so I can smash it on the ideal side and have a clear path to the top.  Even in some cases the hammer ends while the conveyor is going right and I can quickly grab the umbrella and go to the hat and clear the board with 5200 left on the timer.  If I get no pies, I will wait until the conveyor is about to change or descending fireballs which is the standard strategy.

The other day I played an all out type of game which ended at level 13-1 and these were the results:

https://www.youtube.com/watch?v=80Wq6oyDvWQ

http://imgur.com/azwyZ54,QnDt46I,lz2VZZx

For this game I had an average of 9200 points on the pie board with some aggressive play and good hammers.  I decided to pursue this further and play an aggressive game while running boards just to get a bigger and more accurate sample to fit my claims.  This is the link to that game.  It ended at 20-2 at just a million pace.

https://www.youtube.com/watch?v=6D7c0Xlal-A

These were the results on another Jry Chart http://i.imgur.com/QnDt46I.jpg

The average was only around 8600 points, but in my opinion the randomness wasn't entirely favorable in that game.  Since I wanted a bigger example of of pie averages, a Jry chart was created based on both games.  For the average of both games combined, I averaged 8800 points per pie board and the lowest pie board in the entire experiment was 7300 points.  This is the average comparing the experiment and top pie factory averages in full games expressed in this Jry Chart: http://imgur.com/vD3kyfC

Jry math informed me that the average on pie boards in the other sample games were at about 8200-8300 points which would basically be averaging a free pass.  Having about 600 points higher per pie board would approximately result in another 10k for a full game.  After conducting this experiment, I have concluded this will be a advanced strategy I will pursue at reaching my potential for this game.  I would not recommend this strategy to anyone who is in pursuit of their first kill screen, 1 million game. or even 1.1.  It is a very aggressive style that puts you at constant risk and even doing things that aren't normally recommended(Top hammer smashes over the gap).  I would like to hear opinions from other players and maybe give it a try.  If a player like Dean is trying to max out the game, I would recommend giving this a try.  Anyone who has interest in this strategy, I suggest watching every pie board in both games to get an understanding on how I play the board and handle different scenarios.
« Last Edit: August 15, 2015, 12:49:39 pm by lakeman421 »

Offline marinomitch13

  • Spring Jumper
  • *
  • Posts: 1806
    • How to Play DK
    • Awards
Re: Pie Factory Experiment
« Reply #1 on: August 15, 2015, 02:02:51 pm »
I think that is a great strategy. Ethan and I were just discussing the other day how to overhaul the Pie Factory. This seems like a first good plan to try out.

Are you thinking of adding in pie leeching (while waiting) and oil can jumps (to save time in traversing left to right or vice versa) at all? These two things would seem to the be the highest tier in terms of maxing out the board.
"Thou hast made us for Thyself, and our heart is restless until it finds its rest in Thee." -Augustine, Confessions.
Member for 11 Years IGBY 2015 DKF Team Member DK Killscreener Blogger Twitch Streamer

lakeman421

  • Guest
Re: Pie Factory Experiment
« Reply #2 on: August 15, 2015, 03:07:04 pm »
I think that is a great strategy. Ethan and I were just discussing the other day how to overhaul the Pie Factory. This seems like a first good plan to try out.

Are you thinking of adding in pie leeching (while waiting) and oil can jumps (to save time in traversing left to right or vice versa) at all? These two things would seem to the be the highest tier in terms of maxing out the board.

The only time I would add pie leeching into the strategy is when there is 0 chance of getting to the top in the moment(frozen fireballs blocking etc).  If this is the case I look for any points I can grab in the meantime.  Maybe I haven't grabbed the umbrella yet while the fireball is frozen over top right ladder and fireballs remain on the left side.  If I see a pie on the bottom of the board Ill leech that around until things clear up.  The idea behind it would be to limit the damage while waiting which i like to refer as "Meantime points".  They are points I try to grab in the meantime while I cant complete a board or might have an impossible board to complete.  The idea behind it is to grab as many points as I can before a death will possibly occur.  Having a 5k-6k death on a pie board seems a lot better than a 2k-3k death at this point.

To get to your question, I don't see the benefit of leeching pies on the top conveyors since the timer ticks down fast and the inconsistency of spawning pies.  If they consistently came out like on that board in D2K it would probably be a different answer.  There are scenarios though on the top left if the ladder is down and the fireballs are in the way I may wait for a pie to come and and jump it before I try clear the board.  It might give you an extra 100 for the final climb, but it might not either. 

Offline ChrisP

  • Spring Jumper
  • *
  • Posts: 1764
  • I'm going to jump next to your leg.
    • Donkey Blog
    • Awards
Re: Pie Factory Experiment
« Reply #3 on: August 15, 2015, 05:34:12 pm »
We're splitting fireballs on purpose now??

Turns out <Allen> has been ahead of us this whole time...
http://donkeykongblog.blogspot.com

4 Quarters :-* - 800K Avg. Per Qtr. :o - No Restarts 8) - No Proof :'(

7/26/2013   Coin 35,946   710,800   18-1
7/28/2013   Coin 35,947   903,700   22-1
8/16/2013   Coin 35,948   694,100   17-6
8/17/2013   Coin 35,949   893,100   22-1

3,201,700: the $1 World Record?
Member for 11 Years DK Masters - Rank D DKJR Killscreener IGBY 2016 DKF Team Member IGBY 2015 DKF Team Member IGBY 2014 DKF Team Member Blogger Twitch Streamer DK Killscreener CK Killscreener

lakeman421

  • Guest
Re: Pie Factory Experiment
« Reply #4 on: August 15, 2015, 05:52:36 pm »
We're splitting fireballs on purpose now??

Turns out <Allen> has been ahead of us this whole time...
There's a difference in splitting fireballs to get as many points as possible as opposed to splitting the fireballs and still going for the free pass Kappa

Offline up2ng

  • Senior Member
  • *
  • Posts: 242
    • Awards
Re: Pie Factory Experiment
« Reply #5 on: August 16, 2015, 06:43:16 am »
Thanks for sharing Robbie!

This appears to be an excellent idea to look into further for players who have basically mastered the game and are very comfortable and confident with their understanding of fireball behavior to be able to aggressively maneuver around the fireballs in an efficient manner en route to the top, forgoing the possibility of any free pass in an effort to "point press" every Pie Factory.

In the past, that concept was met with a somewhat dubious attitude and it was generally accepted that if you ended up with a board with no bottom pie spawns and fireballs that were staying up top then giving up a clear free pass in order to grab the bottom hammer would likely lose points on average.  At this point I am ready to revisit that assumption and would be curious to see a larger sample of stats using the split fireball strategy for real Pie stages during real games (not from save state). 

Quote
If I get no pies, I will wait until the conveyor is about to change or descending fireballs which is the standard strategy.

I would question this portion of your strategy.  There is a post by Jeff Willms a while back which analyzes the differences in how the board plays out between a relatively early hammer grab and a late (just before reversal) hammer grab in a probabilistic manner and he convinced me after some debate that in most or perhaps all cases there is a higher EV with the earlier hammer grab since it will yield more bottom Pie spawns on average, among other reasons.  Sorry I am not very good about actually finding and posting a link to such threads but perhaps someone else can find it.

At any rate, I did go through those two games that you have posted on Youtube and watched all of the Pie Factory screens.  I am definitely intregued by the overall strategy.  Unfortunately I am dead tired right now and need some rest but I definitely want to think about this further.  When I get a chance, I'd like to post some notes about how a few of the screens played out for discussion.

In general, it appears to me that part of your instinct during these screens is to prioritize fireball smashes over pie smashes.  In my opinion, without looking at this strategy in more depth, I think that this is probably often a mistake and you should look into focusing more on the Pie smashes and only smash the fireballs which encroach a certain threshold of where you are located while smashing these pies.  There are a couple of reasons for this.  First, I think if you stay focused on the pies (and generally chase objects to the right "with" the grain of the conveyor less often) you will end up smashing more objects on average.  The other reason is that pies are higher scoring objects than fireballs when smashed.  (There is another series of threads and/or posts which explain this if these details have been forgotten by the community which perhaps someone can help me find.)  Unless your tactical reasons for clearing the fireballs (in order to enable more efficient Bonus timer preservation, etc) with higher priority exceeds the benefits of staying focused on smashing the pies first, then you may be leaking points with this mentality.  This balance between these two factors (scoring more points vs. preserving more Bonus time) should be further evaluated anyways when developing new strategies.

Anyway, keep up the good work and I hope to continue this discussion.
Donkey Kong:  1,206,800  Kill Screen
Donkey Kong:  898,600     16-5
D2K:                 380,200     L=9
Donkey Kong Junior:  In Progress
Member for 11 Years DK 1.2M Point Scorer Winner of a community event Wildcard Rematch Champion Blogger Former DK Level 1-1 World Record Holder Former DK No-Hammer World Record Holder DK 1.1M Point Scorer Former DK World Record Holder - MAME DK Killscreener DK 1M Point Scorer Individual Board Record Holder Twitch Streamer

lakeman421

  • Guest
Re: Pie Factory Experiment
« Reply #6 on: August 16, 2015, 08:46:08 am »

Quote
If I get no pies, I will wait until the conveyor is about to change or descending fireballs which is the standard strategy.

I would question this portion of your strategy.  There is a post by Jeff Willms a while back which analyzes the differences in how the board plays out between a relatively early hammer grab and a late (just before reversal) hammer grab in a probabilistic manner and he convinced me after some debate that in most or perhaps all cases there is a higher EV with the earlier hammer grab since it will yield more bottom Pie spawns on average, among other reasons.  Sorry I am not very good about actually finding and posting a link to such threads but perhaps someone else can find it.

I think I heard about this in the past from other players although I may not have fully understood it.  Maybe in these cases trying to figure out exactly when to grab the hammer would be another thing to look into.  I kind of just had the mentality that if there was nothing to smash nearby it wasn't worth grabbing, but maybe grabbing it earlier in those situations would be more beneficial based on higher knowledge of the code on that board.

In general, it appears to me that part of your instinct during these screens is to prioritize fireball smashes over pie smashes.  In my opinion, without looking at this strategy in more depth, I think that this is probably often a mistake and you should look into focusing more on the Pie smashes and only smash the fireballs which encroach a certain threshold of where you are located while smashing these pies.  There are a couple of reasons for this.  First, I think if you stay focused on the pies (and generally chase objects to the right "with" the grain of the conveyor less often) you will end up smashing more objects on average.  The other reason is that pies are higher scoring objects than fireballs when smashed.  (There is another series of threads and/or posts which explain this if these details have been forgotten by the community which perhaps someone can help me find.)  Unless your tactical reasons for clearing the fireballs (in order to enable more efficient Bonus timer preservation, etc) with higher priority exceeds the benefits of staying focused on smashing the pies first, then you may be leaking points with this mentality.  This balance between these two factors (scoring more points vs. preserving more Bonus time) should be further evaluated anyways when developing new strategies.

For me this has always just been a survival instinct to get them away from me.  Having a strategy to smash them and get them to a certain side could give me a clear path and a better chance at survival.  With the hammer running out while the conveyor is reversing created awkward situations with the fireballs, which is where was where I struggled.  Smashing the fireballs seemed to be a simple solution for this.  But if we are encouraging to do early hammer grabs every time, than I may feel more comfortable in focusing on just the pies.

I am really excited you came across this thread, because I really wanted to hear your thoughts and possibly ways to improve this strategy.  If you decide to give it a try I would thrilled to see the results.  We will have to continue the scoring more points vs preserving more bonus time experiment and see which one is beneficial.

Offline up2ng

  • Senior Member
  • *
  • Posts: 242
    • Awards
Re: Pie Factory Experiment
« Reply #7 on: August 16, 2015, 03:58:02 pm »
Ok, here is some of the Pie Factory discussion related to the bottom hammer grab that I was thinking about.  Somehow, this information actually ended up landing on my blog pages so it may have been missed by some.  Start from this link and the discussion continues for a few pages:

https://donkeykongforum.net/index.php?topic=286.msg4711#msg4711

The older information related to the fact that pies are higher scoring objects than fireballs I think originated a few years back over on CAGDC, but I think that there is some chance that these threads were migrated here to DKF at some point.  I would have more trouble finding those threads but I hope that someone else knows what I am talking about and tracks that down.  In general, I think it turned out that while (most!) pies have the "standard" 300/500/800 distribution with a 1:2:1 ratio, the fireballs have something more like a 4:2:1 ratio (this might not be exact), leading to a lower EV per smash.
Donkey Kong:  1,206,800  Kill Screen
Donkey Kong:  898,600     16-5
D2K:                 380,200     L=9
Donkey Kong Junior:  In Progress
Member for 11 Years DK 1.2M Point Scorer Winner of a community event Wildcard Rematch Champion Blogger Former DK Level 1-1 World Record Holder Former DK No-Hammer World Record Holder DK 1.1M Point Scorer Former DK World Record Holder - MAME DK Killscreener DK 1M Point Scorer Individual Board Record Holder Twitch Streamer

Offline f_symbols

  • Spring Jumper
  • *
  • Posts: 835
  • wht u mean
    • Awards
Re: Pie Factory Experiment
« Reply #8 on: August 16, 2015, 05:36:36 pm »
...but I hope that someone else knows what I am talking about and tracks that down.  In general, I think it turned out that while (most!) pies have the "standard" 300/500/800 distribution with a 1:2:1 ratio, the fireballs have something more like a 4:2:1 ratio (this might not be exact), leading to a lower EV per smash.

http://www.classicarcadegaming.com/forums/index.php/topic,4164.msg52010.html#msg52010

Member for 11 Years IGBY 2016 DKF Team Member Winner of a community event Former DK Remix World Record Holder DK Remix Killscreener IGBY 2015 DKF Team Member DK 1M Point Scorer IGBY 2014 DKF Team Member Blogger DK Killscreener Twitch Streamer

lakeman421

  • Guest
Re: Pie Factory Experiment
« Reply #9 on: August 17, 2015, 10:54:02 am »
Thanks for all of this information guys.  I have a few days off starting tomorrow, so maybe I'll keep these things in mind while I continue to experiment with these strategies.

Offline YesAffinity

  • Spring Jumper
  • *
  • Posts: 578
    • Awards
Re: Pie Factory Experiment
« Reply #10 on: August 17, 2015, 03:56:26 pm »
Does the flashing 1 up trick help improve the strategy at all?  I notice on many of my L5+ pie factories, fireball 4 and 5 spawn almost on top of each other, using the flashing 1 up trick.  not sure which fireballs you're splitting (sorry if I missed this detail in your posts), but if you're splitting them to have fireball 4 and 5 on the left (at a minimum), using the flashing 1 up trick to prevent the 5th fireball spawn delay will regularly get them on the board more quickly.

Alternatively, timing the flashing 1 up so 1 up is displaying when clearing the ladder board prior to pie factory would increase the possibility of the delayed 5th fireball spawn.  I can't think of a benefit from doing that, but if there's one to be had, I'm sure you gurus will conceive of it.  <popcorn>
Matthew 21:22

DK Arcade PB (verified): 970,200 KS
DK Start PB (verified): 126,600
DK L1-1 PB (verified): 11,400
DK PB 1st Man: 622,000

Donkey Kong Direct Feed How-To - http://donkeykongforum.net/index.php?topic=1413.0
^Now outdated, see instead: http://donkeykongforum.net/index.php?topic=2471.0
Member for 9 Years DK Killscreener Blogger Twitch Streamer

Offline dnickolas

  • Senior Member
  • *
  • Posts: 212
    • Awards
Re: Pie Factory Experiment
« Reply #11 on: June 23, 2016, 08:50:55 pm »
Watched your 1.2 game and noticed that on 6-2 you went to the right for the 4th pie instead of running left after getting the third one. Why? Running left will get a couple others at the edge and you'll be on the left to smash fireballs if they come down causing (imo) easier left spawns.

Is this to put yourself on the right so that a smash causes a right spawn and makes the middle hammer a more likely possibility, or am I wrong and left spawns aren't as sure of a thing as I think when you're on that side?

BTW - nice game.
Member for 8 Years DK Killscreener Blogger Twitch Streamer

Offline dnickolas

  • Senior Member
  • *
  • Posts: 212
    • Awards
Re: Pie Factory Experiment
« Reply #12 on: June 23, 2016, 09:02:55 pm »
At this point I am ready to revisit that assumption and would be curious to see a larger sample of stats using the split fireball strategy for real Pie stages during real games (not from save state)

Why not from save state? I don't see similarities in the way boards play out from saved states versus fresh. Is my MAME messed up, and should I be seeing the same patterns every time?

I understand "real games" being different due to fatigue and nerves, but don't we want to eliminate other factors when examining alternative strategies?
Member for 8 Years DK Killscreener Blogger Twitch Streamer