Beefy Boxes and Bandwidth Generously Provided by pair Networks
The stupid question is the question not asked
 
PerlMonks  

Ranking the Saints by XP Efficiency

by 1nickt (Canon)
on Jan 14, 2016 at 04:24 UTC ( [id://1152739]=monkdiscuss: print w/replies, xml ) Need Help??

Hello fellow monks,

I was curious about my progress in the XP game, and found it hard to visualize so I made this little script to produce a report on "XP Efficiency."

It attempts to produce a ranking of Saints by comparing XP to either number of posts made, or to how long the Saint has been a Monk.

I made a couple of efforts at making the data more relevant. For example, I exclude anyone who hasn't been active for a certain time, and anyone who has not made a minimum number of posts. This eliminates the Monks who have accumulated thousands of points but are not here in the community these days, as well as those who do come by regularly but have never posted anything.

I also try to get at meaningful data by deducting half a point for each day as a Monk, since I'm most interested in XP efficiency based on real contributions to the Monastery, i.e. XP for upvotes and for votes cast, not for votes accumulated through simple longevity. (Of course it is impossible to properly account for all factors, e.g. the fact that there were far fewer Monks in the early days, so posts had fewer potential upvotes they could earn.) Would like to hear any suggestions on how to improve this "normalization" of the data. Mostly though the results appear to be in line with what I would have expected so I don't think it's too far off. Deducting a point per day seems to make a small difference, and only for XP/posts.

It's just for fun, but it helps motivate me to excel as a teacher and as a student. You can run the script and force it to include you in the rankings even if you don't make the top 20 (or whatever you limit the search to), so you can see your progress at a glance.

Usage:

$ perl XP_efficiency.pl Usage: perl XP_efficiency.pl --sort_by (age|posts) Options: --sort_by, -s: Rank monks by XP/posts or by XP/age --limit, -l: Limit results to this many --order, -o: Sort order (asc|desc) --recent, -r: Limit to monks seen in the last n wee +ks --only_include, -i: Skip all monks except this/these one( +s) --force_include, -f: Include this/these monk(s) even if -l + is set --min_posts, -m: Exclude monks with fewer than this ma +ny posts --no_adjust, -n: Don't deduct 0.5 XP for each day of m +onkhood --help, -h: Print this help

A couple of example reports:

$ perl XP_efficiency.pl --sort_by age --min_posts 200 --recent 3 --lim +it 10 --order desc +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 1 2 BrowserUk 159,881 Pope (28) 4,971 36.0644 +22,404 6.9144 | | 2 3 ikegami 127,624 Pope (28) 4,173 34.2347 +19,716 6.2615 | | 3 4 Corion 122,789 Pope (28) 5,783 23.4148 + 9,796 11.9443 | | 4 5 GrandFather 72,446 Sage (25) 3,880 20.4526 + 6,562 10.4489 | | 5 55 Athanasius 24,854 Canon (20) 1,352 20.1077 + 1,458 16.1193 | | 6 664 1nickt 3,115 Curate (13) 214 15.6199 + 468 6.1987 | | 7 42 choroba 29,288 Canon (20) 2,112 14.8885 + 3,739 7.2683 | | 8 20 toolic 43,023 Bishop (22) 3,130 14.7489 + 3,597 11.0906 | | 9 7 tye 70,849 Sage (25) 5,661 13.3272 + 7,887 8.2652 | | 10 60 kcott 24,138 Canon (20) 1,939 13.2484 + 2,407 9.2227 | +--------------------------------------------------------------------- +-----------------+
The next two show the effect of adjusting to discount the daily freebie XP point (not really seen when sorting by XP/age; this is XP/posts):
$ perl XP_efficiency.pl -s posts -m 200 -l 5 -o desc -n +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 1 82 gryphon 18,249 Abbot (19) 5,643 3.7429 + 218 83.7110 | | 2 17 Old_Gray_Bear 46,694 Bishop (22) 4,507 11.9908 + 598 78.0836 | | 3 72 gmax 20,739 Abbot (19) 5,164 4.6478 + 292 71.0240 | | 4 54 Perlbotics 24,920 Canon (20) 3,056 9.4366 + 384 64.8958 | | 5 34 Gavin 32,148 Chancellor (21) 3,599 10.3364 + 600 53.5800 | +--------------------------------------------------------------------- +-----------------+
$ perl XP_efficiency.pl -s posts -m 200 -l 5 -o desc +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 1 17 Old_Gray_Bear 46,694 Bishop (22) 4,507 10.8334 + 598 70.5468 | | 2 82 gryphon 18,249 Abbot (19) 5,643 2.5855 + 218 57.8257 | | 3 54 Perlbotics 24,920 Canon (20) 3,056 8.2794 + 384 56.9375 | | 4 72 gmax 20,739 Abbot (19) 5,164 3.4905 + 292 53.3390 | | 5 34 Gavin 32,148 Chancellor (21) 3,599 9.1792 + 600 47.5817 | +--------------------------------------------------------------------- +-----------------+
Include yourself and someone else and see how you stack up:
$ perl XP_efficiency.pl -s posts -m 200 -l 10 -o desc -f 1nickt -f mer +lyn +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 1 17 Old_Gray_Bear 46,694 Bishop (22) 4,507 10.8334 + 598 70.5468 | | 2 82 gryphon 18,249 Abbot (19) 5,643 2.5855 + 218 57.8257 | | 3 54 Perlbotics 24,920 Canon (20) 3,056 8.2794 + 384 56.9375 | | 4 72 gmax 20,739 Abbot (19) 5,164 3.4905 + 292 53.3390 | | 5 34 Gavin 32,148 Chancellor (21) 3,599 9.1792 + 600 47.5817 | | 6 100 scorpio17 16,117 Abbot (19) 3,253 4.5761 + 323 39.8266 | | 7 61 hsmyers 23,992 Canon (20) 5,452 3.9352 + 520 35.6538 | | 8 44 borisz 27,915 Canon (20) 4,756 5.6351 + 913 25.3658 | | 9 129 dvergin 12,732 Monsignor (18) 5,825 1.3724 + 311 22.2090 | | 10 58 FunkyMonk 24,201 Canon (20) 3,187 7.6302 + 1,024 20.5215 | | 41 6 merlyn 71,157 Sage (25) 5,741 13.1874 + 6,322 10.3474 | | 82 664 1nickt 3,115 Curate (13) 214 15.6198 + 468 6.1987 | +--------------------------------------------------------------------- +-----------------+

Code if you would like to play with it:

#!/usr/bin/perl use strict; use warnings; use feature qw/ say /; use Data::Dumper; $Data::Dumper::Sortkeys = 1; use Getopt::Long; use HTTP::Tiny; use HTML::TableExtract; use Time::Piece; use Number::Format qw/ format_number /; use List::Util qw/ any /; my ( $sort_by, $order, $limit, $recent, $only_include, $force_include, + $min_posts, $no_adjust ); GetOptions( 'sort_by|s=s' => \$sort_by, 'order|o=s' => \$order, 'limit|l=i' => \$limit, 'recent|r=i' => \$recent, 'only_include|i=s@' => \$only_include, 'force_include|f=s@' => \$force_include, 'min_posts|m=i', => \$min_posts, 'no_adjust|n' => \$no_adjust, 'help|?' => \&print_usage, ); minor_validation(); my $url = 'http://perlmonks.com/?node_id=3559'; my $headers = ['#', 'User', 'Experience', 'Level', 'Writeups', 'User S +ince', 'Last Here']; my $data; fetch_data( $url ); print_data( $data ); exit; ###################################### sub fetch_data { my $url = shift; my $page = HTTP::Tiny->new->get( $url ); die 'Died. No response' if not $page; die "Died. $page->{'status'} $page->{'reason'}" if not $page->{'su +ccess'}; my $table = HTML::TableExtract->new( 'headers' => $headers ); $table->parse( $page->{'content'} ); my $now = time(); foreach my $row ( $table->rows ) { process_row( $now, $row ); } } sub process_row { my $now = shift; my $row = shift; my ( $rank, $monk, $xp, $level, $posts, $created, $last_here ) = @ +{ $row }; return if $last_here =~ /year/ or ( $last_here =~ /(\d+) week/ and + $1 > ( $recent || 4 ) ); return if $sort_by eq 'posts' and $posts eq 'None'; return if $min_posts and ( $posts eq 'None' or $posts < $min_posts + ); return if $only_include and not any { $_ eq $monk } @{ $only_inclu +de }; $created = Time::Piece->strptime($created, '%Y-%m-%d %R')->epoch; my $age = $now - $created; my $age_days = int( $age / 86400 ); my $xp_adjusted = $no_adjust ? $xp : $xp - ( $age_days / 2 ); my %record = ( rank => $rank, xp => $xp, level => $level, posts => $posts, by_posts => $posts eq 'None' ? 0 : ( $xp_adjusted / $posts ), age => $age_days, by_age => ( $xp_adjusted / $age ) * 100000, # for scale ); $data->{ $monk } = \%record; } sub print_data { my $data = shift; $sort_by = 'by_' . $sort_by; my %sortable = ( map { $_ => $data->{ $_ }->{ $sort_by } } keys %{ + $data } ); my @results; foreach my $monk ( sort { $sortable{$a} <=> $sortable{$b} } keys % +sortable ) { $data->{ $monk }->{'posts'} =~ s/None/0/; push @results, { rank => $data->{ $monk }->{'rank'}, monk => $monk, xp => format_number( $data->{ $monk }->{'xp'} ), level => $data->{ $monk }->{'level'}, age => format_number( $data->{ $monk }->{'age'} ), by_age => format_number( $data->{ $monk }->{'by_age'}, 4 +, 1 ), posts => format_number( $data->{ $monk }->{'posts'} ), by_posts => format_number( $data->{ $monk }->{'by_posts'}, + 4, 1 ) }; } @results = reverse @results if $order and $order eq 'desc'; print <<EOT; +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ EOT my $count = 0; foreach my $result ( @results ) { ++$count; if ( $limit and $count > $limit ) { last if not defined $force_include; next unless any { $_ eq $result->{'monk'} } @{ $force_incl +ude }; } format = @ @### @#### @<<<<<<<<<<<< @>>>>>> @>>>>>>>>>>>>>>> @>>>>>> @>>>>>>> @ +>>>>>> @>>>>>>>> @ '|', $count, $result->{'rank'}, $result->{'monk'}, $result->{'xp'}, $r +esult->{'level'}, $result->{'age'}, $result->{'by_age'}, $result->{'p +osts'}, $result->{'by_posts'}, '|' . write; } print '+---------------------------------------------------------- +----------------------------+'; } sub minor_validation { print_usage() if ( ( ! $sort_by or $sort_by !~ /age|posts/ ) or ( $order and $order !~ /asc|desc/ ) ); } sub print_usage { say <<"EOT"; Usage: perl $0 --sort_by (age|posts) Options: --sort_by, -s: Rank monks by XP/posts or by XP/age --limit, -l: Limit results to this many --order, -o: Sort order (asc|desc) --recent, -r: Limit to monks seen in the last n wee +ks --only_include, -i: Skip all monks except this/these one( +s) --force_include, -f: Include this/these monk(s) even if -l + is set --min_posts, -m: Exclude monks with fewer than this ma +ny posts --no_adjust, -n: Don't deduct 0.5 XP for each day of m +onkhood --help, -h: Print this help EOT exit; } __END__

edit: reduce XP by only 0.5 per day; switched to Time::Piece as DateTime::Format::Strptime has compile problems on Strawberry Perl; switched to format_number() because round() does not provide trailing zeroes; increase XP/Age by two more orders of magnitude, for scale


The way forward always starts with a minimal test.

Replies are listed 'Best First'.
Re: Ranking the Saints by XP Efficiency
by choroba (Cardinal) on Jan 14, 2016 at 09:36 UTC
    ++ Isn't there a place to gather these? I already have at least two related posts:

    Update: If you don't want to install Number::Format and DateTime::Format::Strptime, you can use Time::Piece (in core since 5.10) and sprintf:

    use Time::Piece; # ... $created = 'Time::Piece'->strptime($created, '%Y-%m-%d %R')->epoch +; # ... xp => sprintf('%d', $data->{$monk}->{xp}), level => $data->{$monk}->{level}, age => sprintf('%d', $data->{$monk}->{age}), by_age => sprintf('%.4f', $data->{$monk}->{by_age}), posts => sprintf('%d', $data->{$monk}->{posts}), by_posts => sprintf('%.4f', $data->{$monk}->{by_posts}), # ...
    ($q=q:Sq=~/;[c](.)(.)/;chr(-||-|5+lengthSq)`"S|oS2"`map{chr |+ord }map{substrSq`S_+|`|}3E|-|`7**2-3:)=~y+S|`+$1,++print+eval$q,q,a,
      thanks choroba, i hope the OP is updated with your patch because that DateTime::Format::Strptime wants a newer version of Params::Validate that does not install in my strawberry perls.

      L*
      There are no rules, there are no thumbs..
      Reinvent the wheel, then learn The Wheel; may be one day you reinvent one of THE WHEELS.
Re: Ranking the Saints by XP Efficiency
by ww (Archbishop) on Jan 14, 2016 at 15:47 UTC

    I find your explanation of the deduction for days-as-a-monk

    "I also try to get at meaningful data by deducting a point for each day as a Monk, since I'm most interested in XP efficiency based on real contributions to the Monastery, i.e. XP for upvotes and for votes cast, not for votes accumulated through simple longevity."
    a bit curious for two reasons:

    1. deducting one "point for each day as a Monk" seems a less than rigorous way of accounting for this (from Voting/Experience System):
      Visiting the site regularly.
          You have 25% chance of gaining 2 XP points once a day if you were logged in within the past 24 hours
          from when the vote fairy does the rounds.

      Would it not be closer (albeit, still imprecise since as you note there are numerous variables for which we can't readily account) to deduct 0.5 points/day in the expectation that over time ((2 points * 25%) * days)) would be statistically more probable (at least for a moderately regular visitor)?

            and
       
    2. as to XP/node, would it not more clearly reflect a merit ranking for a users posted nodes by using the data from Perl Monks User Search. Obviously that's perhaps a bit more than a SMOP ... more like QLOP ("quite'a lot of programming") because one would have to add code to
      1. ascertain the actual number of writeups by each user
      2. insert a user name and the number of writeups into ("Show") limit in PM User Search
      3. extract and process that data... for as many cases as specified on the CLI

    And even that has a major flaw: PM User Search will report XP ONLY FOR THOSE NODES UPON WHICH THE USER BEING ASSESSED IS THE LOGGED-IN USER. There are probably other defects in this set of comments, which is really inspired only by admiration for the OP! (which does actually need a less problematic formatting module, as noted by choroba, above).

    UPDATE: re 2. above: I should be clear: that's NOT OP's stated intent; it's merely a variant approach that I would happen to favor, were it feasible.
    and, way too many fixes of bad markup, punctuation, etc. My Bad!

      Thanks, ww. I was lazy and didn't reread the Voting/Experience System explanation.

      Script changed to only deduct 0.5 XP per day as a monk. This makes a meaningful difference in score to certain Monks who were lurkers for a long time and later became active posters -- but that doesn't guarantee an upward change in ranking since everyone else also benefits from the change.

      Before:

      +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 134 nnn CertainMonk nnnnn Vicar (15) 4,963 0.3876 + 791 2.1024 | +--------------------------------------------------------------------- +-----------------+
      After:
      +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 139 nnn CertainMonk nnnnn Vicar (15) 4,963 0.9662 + 791 5.2396 | +--------------------------------------------------------------------- +-----------------+

      The way forward always starts with a minimal test.
Re: Ranking the Saints by XP Efficiency
by chacham (Prior) on Jan 14, 2016 at 14:23 UTC
Re: Ranking the Saints by XP Efficiency
by toolic (Bishop) on Jan 14, 2016 at 20:47 UTC
    I don't think all posts are created equal. For example, I'd be willing to bet that the number of votes a post receives is a strong function of the depth of the post in a thread. My guess is that Re^7: posts receive very few votes as compared with root nodes and Re: replies. If a monk has a tendency to post deep into a thread, that monk's XP/Posts number would be lower that a monk who tends not to dive deep.

    It would be interesting to see the average Rep for posts as a function of depth for a given monk. Since this information is not available on the Saints in our book page, this would be more compute-intensive exercise (not to mention that the typical monk can not see another's Rep data).

      While that's true, the OP's stated intent is to determine the "XP efficiency" of the various monks. If deeper posts tend to receive fewer votes (and I agree that this is almost certainly the case), then that just means that making deep posts is "inefficient". Treating deep (or late, for that matter) replies differently would only serve to mask that inefficiency, which seems counter to the OP's purpose.

      Yes.

      Other factors include time of day, weekend or not -- the longer a thread is "recent" the more views and therefore votes it will get -- and probably several more I can't even think of.

      It's definitely not a scientific tool, but the law of averages makes it somewhat useful as a progress bar, I think. For example, the depth-of-post factor should apply to all Monks roughly equally, if we assume that all Monks behave the same in terms of how deep into a thread they would be likely to post.


      The way forward always starts with a minimal test.
        if we assume that all Monks behave the same in terms of how deep into a thread they would be likely to post
        Agreed, but I tend to tire easily, whereas BrowserUk tends to be tenacious. My guess is his XP/Posts would crush mine if deep posts were excluded:
        +--------------------------------------------------------------------- +-----------------+ | Pos | St. | Monk name | XP | Level | Age | XP/Age | +Posts | XP/Posts | +--------------------------------------------------------------------- +-----------------+ | 1 2 BrowserUk 159,881 Pope (28) 4,971 36.0644 +22,404 6.9144 | | 8 20 toolic 43,023 Bishop (22) 3,130 14.7489 + 3,597 11.0906 |

Log In?
Username:
Password:

What's my password?
Create A New User
Domain Nodelet?
Node Status?
node history
Node Type: monkdiscuss [id://1152739]
Approved by Corion
Front-paged by ww
help
Chatterbox?
and the web crawler heard nothing...

How do I use this?Last hourOther CB clients
Other Users?
Others perusing the Monastery: (5)
As of 2024-04-19 13:49 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    No recent polls found