Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

51

Goko Dominion Online / Re: Salvager reports incorrect ISO level

« on: September 29, 2014, 06:53:01 pm »

Where can we post when Salvager is reporting incorrect ISO levels? Mine is working fine, but I've noticed some players aren't.

Yoshiball (JPN) is always at -9 when actually 33
Ok-2- has changed goko rating from 4600 to 5300 and is always shows lev 16
Light Yagami is always lev 15 when really 27
TOTO_TOTO is reported as -12 by Salvager when really 12

I suppose I can just make mental notes of these, but fixed ratings might help them get more games.

This happens when people change their usernames.  It's because some parts of the project use player hash (as they should) as a unique id while other parts still use player name (which is horrible).

I've just been hacking the code for people who notice and ask me to fix it specifically for them, but the right solution is to make the database do everything with player hash.  It's a fairly involved problem though, because the logs only record player names.  I think it's been discussed in the Isotropish thread somewhere.

52

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 29, 2014, 12:06:50 pm »

I just signed up Salvager on bountysource.com.  You can use it to offer and request cash for implementing specific features.  It seems to hook in with github rather nicely:

https://www.bountysource.com/trackers/570746-goko-salvager

So if you want a feature, you can offer cash for it.  If you're willing to implement a feature, you can specify a price that you'll do it for.

As a test, I just created a $5 "bounty" for the request that TheMirrorMan a few months back: "Add the word "beta" in case of a beta version to the title screen."  So now whoever implements that feature can go and claim that $5, assuming I understand this right.

This might be a good solution for:
1) People who have had outstanding feature requests for months that nobody is doing
2) People who would like to implement features but have more important things to be doing
3) People who have asked for a way to donate to the project

53

Goko Dominion Online / Re: Log Search Engine

« on: September 29, 2014, 08:14:07 am »

Anyone else with HTML and/or Python skills wants to look into it?
https://github.com/aiannacc/goko-dominion-tools/blob/master/gdt/logsearch/logsearch_handler.py
https://github.com/aiannacc/goko-dominion-tools/blob/master/web/logsearch.html

I don't have either skill, really, so feel free to ignore this.  But it looks to me like the self.get_argument in line 53 defaults to stripping white space from the beginning and and of the string.  Certainly this behavior is what I see.  I can add whitespace to the front and rear of my name and still find all my games.

Line 53:

Code: [Select]

arg_str[p] = self.get_argument(p, default=self.search_params[p])
This doesn't actually strip any whitespace directly, though the array name arg_stp sort of suggests that it might.

However, the self object is a RequestHandler from the third-party tornadoweb package I use.  And if you look at the package's srouce code, its get_argument() method delegates to an internal method:

Code: [Select]

    def _get_arguments(self, name, source, strip=True):
        values = []
        for v in source.get(name, []):
            v = self.decode_argument(v, name=name)
            if isinstance(v, unicode_type):
                # Get rid of any weird control chars (unless decoding gave
                # us bytes, in which case leave it alone)
                v = RequestHandler._remove_control_chars_regex.sub(" ", v)
            if strip:
                v = v.strip()
            values.append(v)
        return values

... and the v = v.strip() part does in fact strip whitespace!  So I don't know whether you misread my code or went the extra mile to find the problem, but either way thanks!

I implemented the fix, which was just a one-liner once I knew the problem.  Searching for "Holger " works properly now.

54

Goko Dominion Online / Re: Log Search Engine

« on: September 28, 2014, 03:40:10 pm »

You're listed as "Holger " with a space all throughout my database (games, rating, playerinfo).  So Goko definitely thinks there's a space there.

That's strange. I can try and remove the space from my username, if that helps.

Yes, that would work.

55

Goko Dominion Online / Re: Log Search Engine

« on: September 28, 2014, 02:59:50 pm »

I never changed my username, AFAIR. There shouldn't be a space at the end of my username, but how could I check if there is? Goko wouldn't show a space, and it also wouldn't show a non-capital "H" since the usernames are given only in capital letters. The "H" does seems to be a pixel taller than the other letters of the name, but it's hard to tell...
I don't know if there's another "holger", either; I don't remember if I actually played those New Year's Day games or not, unfortunately.

On Isotropish, my games are given correctly under "Holger" (around level 39 currently) - but I wouldn't see a space at the end of the username either, obviously. I suppose you would?

You're listed as "Holger " with a space all throughout my database (games, rating, playerinfo).  So Goko definitely thinks there's a space there.

The search for ought to be able to handle a space at the end of the name, but something appears to be eating it.  I can't figure out why, but I'll try again later.

Anyone else with HTML and/or Python skills wants to look into it?
https://github.com/aiannacc/goko-dominion-tools/blob/master/gdt/logsearch/logsearch_handler.py
https://github.com/aiannacc/goko-dominion-tools/blob/master/web/logsearch.html

56

Goko Dominion Online / Re: Log Search Engine

« on: September 28, 2014, 02:41:00 am »

If you've been playing pro games since 7/14 but they're not in the database, then something else is going on.  It may be on MF/Goko's side, as sometimes they just fail to create a log and there's nothing gokosalvager.com can do about it.  But if you:
1. play a pro game
2. and the log exists
3. and the game doesn't show up in the log search

... then please post links to the log and the failed search here and I'll try to track down the bug.

The engine doesn't find any games that I played since (or before) New Year's Day; actually I'm not quite sure if I played those 5 games at all, but I played plenty since then.
I played mainly pro games and adventure games against bots, but not exclusively. E.g. this game can be found by searching for my opponent, but not under my name. The games also usually caused my Isotropish rating to update.
In case it matters, I never changed my username, only my computer. The few games shown give my name as "holger", while Isotropish uses "Holger" for whatever reason.

Edit: I recall the search engine working fine for me several months ago; unfortunately I don't remember if this was before or after January.

You, sir, are a sneaky one.  But I figured out your game:

Code: [Select]


 2012-12-08 10:34:20    "Holger "
 ...
 2013-12-18 05:02:42    "Holger "
 2013-12-31 01:17:26    "holger"
 ...
 2014-01-01 13:19:21    "holger"
 2014-01-02 11:46:15    "Holger "
 ...
 2014-09-26 11:18:48    "Holger "


So does your name end in a space, or is there something wrong with my log parsing?  And is "holger" a different player, or did you change your username for the 2014 new year?

57

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 25, 2014, 05:15:23 pm »

I like the extension very much and the new buttons are cool, but that was so sad. The buttons which should protect me from misclicking caused my worst misclick. Maybe there is a way to change this problem.

That is weird... typing in the chat window ought to make it grab focus, but it looks like focus stayed on the End Turn button (presumably from a previous turn where you used it).  I notice you're using the Goko default chat window instead of the sidebar chat, that may be related.  What OS and browser were you using?  I don't like that behavior either, but if I can replicate it then I can hopefully change it.

Maybe the events didn't get processed in the right order?  Say the listener that triggers the button is ahead of the listener that grabs focus for chat.

To be continued in github: https://github.com/aiannacc/Goko-Salvager/issues/249

58

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 20, 2014, 12:41:08 pm »

Oh man... I didn't even notice this.  I'd added a filter to my email that seems to have caught all the Salvager stuff posted here.

Fixing the SSL certificate now...

New SSL certificate installed.  You may need to reload to get everything working again.  Sorry about that.

59

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 20, 2014, 12:21:20 pm »

Oh man... I didn't even notice this.  I'd added a filter to my email that seems to have caught all the Salvager stuff posted here.

Fixing the SSL certificate now...

60

Goko Dominion Online / Re: Update 28 August 2014

« on: September 15, 2014, 08:26:15 pm »

Normally you get 15 coins for a win, 10 for a tie.  Perhaps it was graded as a tie for some strange reason? When you tie, it still says "you're the winner".

No, you only get 10 coins for a win against a bot (I think you get 15 coins only for wins against human opponents).
A tie would explain the rating loss, though I'd expect it to be even more than 30 points then.

I am now confident it was due to increased uncertainty after not playing for several days that carried the loss. Evidently displayed rating only changes when you play a game while true rating changes daily . Thus if you don't play for a few days a dozen or more point loss (30 here) will be incorporated into your next game.

This explanation requires the player to have initial uncertainty of ~1400.  For a regular player, uncertainty is < 300.  At +1% uncertainty per day, that's an awfully long time sitting inactive, about 5 months without a game.

Code: [Select]

Initial Uncertainty = 1400:
    AI Rating - Before:
    6500.00 +/- 1400.00
    Displayed as 3700
   
    AI Rating - After:
    6513.54 +/- 1390.26
    Displayed as 3733
   
    Displayed Change: -33

Python script to produce this output is attached, including rating info for me and Serf Bot.  The trueskill Python module is required.

61

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 13, 2014, 07:48:18 pm »

Not sure if this is the right thread to report things related to the tools on gokosalvager.com, but a few things don't seem to be updated for Prince.  Kingdom Visualizer gives a blank page for the Prince games, searching logs for games with Prince gives "Error parsing supply" and searching for games without Prince is another blank page.

Thanks.  Fixed.

62

Goko Dominion Online / Re: Update 28 August 2014

« on: September 11, 2014, 01:58:23 pm »

Guys, I'm really sorry.  My earlier post about the changes to the rating system was completely wrong.  I had a sneaky Python bug that left the code running without visible errors, but which produced all the wrong numbers.

I have no reason to think the rating system has changed.  The rating changes I'm getting still jive with Goko Pro as I described it back in April.  As for my -30 game against Serf Bot, I still have no idea how that happened.

(Better Q, why aren't they paying you, but I think that's a different discussion ).

Maybe this is why.

63

Goko Dominion Online / Re: Update 28 August 2014

« on: September 11, 2014, 12:27:28 pm »

Can you talk to goko directly?  Why don't this listen to you?

(Better Q, why aren't they paying you, but I think that's a different discussion ).

I've tried it on a few occasions.  I mostly just get ignored.  I sent this to a developer on MF a few days before I posted their rating algorithm.  I never got a response.



I've blacked out the developer's name and the quote from one of their forum posts.  No point in picking out an individual when it's the institution that's at fault.

64

Goko Dominion Online / Re: Update 28 August 2014

« on: September 11, 2014, 11:56:31 am »

If you played Serf Bot, it's quite likely that you'd lose points for a win due to an increased uncertainty; against other bots, it's less likely but not impossible if you have very low uncertainty.

Edit: I forgot, Goko was "perverting" the rating system to "hide" any point losses after a win. So have they just returned to unmodified TrueSkill now, or have they also changed their TrueSkill parameters for the worse?

A -30 change is way too big.  Under the old system, a 7000 player might have lost something like 4 points of displayed rating by beating Serf Bot.  This is something different.

The better question is this:

They've already said they're essentially rewriting the whole thing from scratch.  Why even think about discussing the possibility of making these sorts of changes?

Ah.  Yes, it seems that they've changed the system.  I just played a test game and got some pretty crazy results:

Edit: Ack!  No, I'm completely wrong.  I wrote a bug into my code that screwed up all my numbers.  Here's my original incorrect post:

Estimating TS parameters (beta, tau, draw_prob) using Game 1...
Error minimization finished.
Residual Error: 41.8109

Initial ratings:
 A: 5995.29 +/- 279.27
 B:  963.13 +/- 1344.96
Expected post-game ratings:
 A: 5995.32 +/- 279.26
 B:  962.48 +/- 1343.81
Observed post-game ratings:
 A: 5996.90 +/- 280.42
 B:  926.24 +/- 1323.07
Rating prediction error: 1748.1527

First, my optimizer can no longer find the TS parameters (residual error should be zero).  Second, the old Goko TS parameters don't predict the rating changes correctly anymore.  Third, the new results are insane.

This was a game between a 5400 player and a 0 player and I resigned immediately as the 0 player.  Under the old Goko Pro system, the 0 player (actually 963.13 +/- 1344.96) would have lost almost nothing.  Instead, he dropped 37 points of mu.  The 5400 player (actually 5995.29 +/- 279.27) also gained a "lot" more than he would have under the old system.  Before, he only would have moved to mu = 5995.32, a gain of 0.03.  Instead he went to 5996.90 for a gain of 1.61.

In brief, both players moved by about 57x as much as they would have in the old days.  So either is WAY more volatile, or it's doing something pretty weird for players with very different ratings.  And since my optimizer no longer works, it's either it's no longer pure TrueSkill or their new parameters are REALLY out there.

So here's my pure, uninformed guess: They're still running TS, but they've tweaked it.  Those tweaks are either buggy or they've only tested them on similarly-rated players, so we're getting some crazy results when the rating differences are large.

Anybody want to collect some data for me?  I need ratings for both you and your opponent before (or during) and after the game.  Open a JS console and run this command, using your own player names instead of "guest4261":

conn.getPlayerId({developerId:'', playerName:'guest4261'}).then(function(x) { conn.getRating({
      version: 1,
      playerId: x.playerId,
      ratingSystemId: mtgRoom.options.ratingSystemPro
  }).then(function (resp) {
      console.log(resp.ratingData);
  });
})

And here's the output now that I've fixed my code.  It looks like the rating system hasn't been changed.

Code: [Select]

Estimating TS parameters (beta, tau, draw_prob) using Game 1...
Error minimization finished.
Residual Error: 0.0000

Error-minimizing Parameters:
beta:      1375.00
tau:         27.50
draw_prob:    0.05

Initial ratings:
 A: 5995.29 +/- 279.27
 B:  963.13 +/- 1344.96
Expected post-game ratings:
 A: 5996.90 +/- 280.42
 B:  926.24 +/- 1323.07
Observed post-game ratings:
 A: 5996.90 +/- 280.42
 B:  926.24 +/- 1323.07
Rating prediction error: 0.0000

65

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 11, 2014, 12:35:05 am »

Cool! Now to get something like this incorporated into an "Advanced" mode of the Salvager Kingdom Generator...

I've added AI's "Formula for modified card frequencies" to the kingdom generator and submitted a pull request.

https://github.com/aiannacc/Goko-Salvager/pull/248

Edit: I should mention that in the kingdom generator interface, when adding cards to the more/less/exclude lists, enter the card names separated by + or ,

Now available in the current beta (v2.5.5.0).  Thanks, michaeljb!

66

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 11, 2014, 12:34:22 am »

the extension isn't working properly for me today, i can see the log and VP counter, but can't scroll up/down the log, or click the end actions/end turn buttons or reply to the chat. Wondered if anyone else had had a similar problem ( I had it 2 days back as well). Using Windows 8, newest Chrome and Salvager 2.5.4.9 (the beta version).

There doesn't seem to be an epidemic, so it's probably something specific to your machine.  Got any uncommon extensions that might be fighting with Salvager?  Does Goko work properly for you with the extension off?  Post a screenshot and your Javascript log and I'll look for the error.

67

Goko Dominion Online / Re: Formula for modified card frequencies

« on: September 11, 2014, 12:30:42 am »

michaeljb's implementation is now in the beta (v2.5.5.0).  Here's what it looks like:

68

Goko Dominion Online / Re: Update 28 August 2014

« on: September 11, 2014, 12:13:17 am »

Play vs. Bots now uses the Pro rating system, which is used for fair games. If you want to play bots under the Casual system, you still can by creating a table with the casual system and inviting a bot.

Hmm... so the bad news that now I have to actually play out the game when I want to test a feature.  The other bad news is that I lose 30 points even for a win:



I'm pretty sure this shouldn't be possible unless someone at Goko has been messing with their TS parameters.  And they'd have to be pretty insane parameter at that.  Is this just a bug?

69

Goko Dominion Online / missed opportunity

« on: September 10, 2014, 07:34:00 pm »

Just found this in my e-mail.  Too bad it's already past 7 PM. 

ANDREW,
I have an opening today to get GOKOSALVAGER instantly qualified for up to $500,000 of unsecured funds.  Please give me
a call before 7:00 pm EST today so I can get you in the system.  Over 90% of applicants qualify for this special program, but I need to
submit you application today.
Thanks, and I look forward to you assisting you!.
Best Regards,
Ben Weston
Senior Funding Specialist
Merchant Cash Advances
866-283-1958

70

Goko Dominion Online / Re: Goko's Rating System, Part 2: Reverse Engineering

« on: September 10, 2014, 09:31:43 am »

I also hope that AI will still get around to posting "Part 3". I'm not quite certain if Goko's ranking (minus the bugs) is worse than Isotropish's; Isotropish uses an extremely high initial uncertainty, making the board very conservative wrt new players...

A couple-hundred games tends to drop a new player into the normal uncertainty range.  For example, here are the top active players with fewer than 300 games.  Most veteran players end up with an uncertainty (3*sigma) around 10.

      pname      |   mu    | 3*sigma | numgames | level
-----------------+---------+----------+----------+-----------
 awall           | 61.3037 |  15.2337 |      129 |   46.0700
 Marin           | 54.5226 |  11.6589 |      173 |   42.8635
 Jean-Michel     | 50.8940 |  10.7637 |      249 |   40.1303
 DG              | 59.1405 |  19.8219 |       46 |   39.3186
 Holger          | 50.2388 |  11.5032 |      236 |   38.7356
 Drab Emordnilap | 49.7575 |  11.7360 |      172 |   38.0214
 Wisper          | 47.9380 |  11.4375 |      191 |   36.5004
 loppo           | 45.3395 |  10.4670 |      295 |   34.8726
 Young Nick      | 45.6504 |  10.8579 |      224 |   34.7927
 GeoLib          | 44.7361 |  10.9362 |      210 |   33.7999
 Käkkäräfasaani  | 46.4833 |  12.8319 |      133 |   33.6513
 Shinigami       | 43.3002 |  10.4169 |      298 |   32.8833
 Madman          | 44.2269 |  11.6328 |      201 |   32.5940
 Simon (DK)      | 45.5164 |  14.3709 |      104 |   31.1455
 SawneyBean      | 41.6755 |  10.7028 |      247 |   30.9726
 Yaju            | 50.4640 |  19.9638 |       49 |   30.5000
 mpsprs          | 43.9237 |  13.5030 |      112 |   30.4206
 BeeeeJ          | 41.3244 |  10.9389 |      227 |   30.3855
 MrFrog          | 41.3698 |  11.4582 |      216 |   29.9117
 Trevor Pasanen  | 41.6197 |  11.9886 |      166 |   29.6312

Also bear in mind that brand new players often really are changing their skill level, which affects how quickly their uncertainty drops.

IMO, Isotropish's bigger weakness is how long it takes to figure out that a veteran player is improving, rather than just having a lucky streak.  More on this shortly... I really am going to post part 3.  I've been sitting on a near-finished version for quite a while.

71

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 08, 2014, 06:37:07 pm »

feature request: a friends list.

Third time's the charm

It's not that we're ignoring you.  It's just that nobody has really done any Salvager development in the last few months.

72

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 08, 2014, 03:55:54 pm »

At the risk of being overly glib, note that it's only in ~0.03% of games that this even has the potential to affect VP counts.

Twice as high as that, I think, since it could affect Fairgrounds as well as Gardens, if the counter thinks you've spent your last Spoils but you haven't.

ThaddeusB gets +1 for noting that Pillage also produces Spoils.  The correct frequency is 0.05%.

Alan gets +1 for noting Fairgrounds, but -2 for not noticing that I'd already counted it.

First person to note Black Market and Young Witch gets... to do the math themselves.

73

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 08, 2014, 08:17:15 am »

Actually, it may not be possible to solve this in all cases.  You can't tell from just the log whether it was Counterfeit played on 1 Spoils or Counterfeit on nothing followed by playing 2 Spoils.

Seems solveable in most cases, except some weird ones, but you'd have to track the contents of the player's hand (can you do that during the game, or only with the post-game log that says which 5 cards you drew at end of turn?).  The deck can't lose more Spoils than were in the player's hand.  One edge case where things are still ambiguous is if a player has two Spoils in hand, double-plays one with Counterfeit, then declines to play the other.  There might be other weird cases involving multiple Counterfeits.

Yes, you can handle many cases by adding hand-tracking logic to the post-game log viewer.  That's no good for Salvager's in-game log viewer and VP counter though.  Further discussion on github: https://github.com/aiannacc/Goko-Salvager/issues/247

At the risk of being overly glib, note that it's only in ~0.03% of games that this even has the potential to affect VP counts.

74

Goko Dominion Online / Re: Goko Dominion Salvager Discussion

« on: September 07, 2014, 11:03:59 pm »

VP counter bug (also effects log viewer): See this game - http://gokosalvager.com/static/logprettifier.html?20140907/log.53be12a0e4b00ccce8075fbc.1410124895271.txt  - The salvager and log viewer both say my opponent had 32 when he actually had 35.  Most likely, it is miscalculating the Gardens value for some reason.

Ah... it's the Counterfeit-Spoils interaction:

Code: [Select]

PeWi plays Counterfeit
PeWi plays Spoils
PeWi plays Spoils

When either Salvager or the log viewer sees this, it thinks that 2 Spoils were played and returned to the Supply, when really it was only 1 Spoils played twice.  So it's getting the deck counts wrong.

Actually, it may not be possible to solve this in all cases.  You can't tell from just the log whether it was Counterfeit played on 1 Spoils or Counterfeit on nothing followed by playing 2 Spoils.

75

Goko Dominion Online / Formula for modified card frequencies

« on: September 05, 2014, 03:52:28 pm »

For casual and unrated games, SCSN and I like to modify the frequency with which particular cards are chosen.  We like to play games with Black Market or Tactician, but we're not so big on the sort of game that IGG or Cultist leads to.  As for the cards that are either very weak or very strong and randomizing, it's nice to see them on occasion but maybe not as often as in full random.

Here's the formula that generates modified frequencies we like:

Code: [Select]

10 ( 9*(Black Market + Procession + Hermit + Inn + Watchtower + Mint + Tactician + Bishop + Prince)
+ 3*( ALL / (Black Market + Procession + Hermit + Inn + Watchtower + Mint + Tactician + Bishop + Prince)
          / (Pirate Ship + Thief + Transmute + Adventurer + Scout + Coppersmith + Harvest + Cache + Noble Brigand + Taxman + Rebuild + Minion + Tournament + Possession + Pearl Diver + Swindler)
          / (Ill-Gotten Gains + Cultist + Fool's Gold) )
+ 1*(Pirate Ship + Thief + Transmute + Adventurer + Scout + Coppersmith + Harvest + Cache + Noble Brigand + Taxman + Rebuild + Minion + Tournament + Possession + Pearl Diver + Swindler) )
This formula:

• excludes (IGG, Cultist, and FG) entirely
• chooses (Black Market, Procession, etc.) 3x as often as normal
• chooses (Pirate Ship, Thief, etc.) and (Rebuild, Tournament, etc) 1/3 as often as normal.

We find this tends to generate a wider range of strategies than full random.  We often find ourselves saying "well that was a weird game."

The same technique can bias the kingdom towards engines (add villages to the 3x list) or away from attacks (add them to the 1/3x list).  What's nice is that you don't have to completely force or eliminate those cards like "Village, 9 * (ALL / Attack)" would.

To emphasize different cards or change their frequencies, you can tweak the script that generates the formula:

Code: [Select]

#!/bin/bash

# Choose these cards 300% as often as normal
more="(Black Market + Procession + Hermit + Inn + Watchtower + Mint + Tactician + Bishop + Prince)"

# Choose these cards 33% as often as normal
less="(Pirate Ship + Thief + Transmute + Adventurer + Scout + Coppersmith + Harvest + Cache + Noble Brigand + Taxman +  Rebuild + Minion + Tournament + Possession + Pearl Diver + Swindler)"

# Never choose these cards
never="(Ill-Gotten Gains + Cultist + Fool's Gold)"

# Choose all other cards at the normal rate
normal="( ALL / $more / $less / $never )"

# Note: The kingdom generator can't specify non-integer rates (e.g. 1/3 for the "less" cards).
#       Instead, we can multiply all the rates by 3 so that they're all integers.  The kingdom
#       generator will effectively reverse the scaling.  That is
#
#       300% --> 9x for the "more" cards
#        33% --> 1x for the "less" cards
#         0% --> 0x for the "never" cards
#       100% --> 3x for the "normal" cards (everything else)
#
echo "10 ( 9*$more + 3*$normal + 1*$less )"

If you don't have a POSIX shell, you can use a site like this one to run it.  Changing which cards that it draws more or less often is straightforward, but changing the frequencies is a little trickier.  I've given very brief explanations in the script comments, but feel free to ask for help here or in PM.