diff --git a/package.json b/package.json
index fd5f3a31de..f60ceb65d5 100644
--- a/package.json
+++ b/package.json
@@ -16,9 +16,7 @@
     "guid": "*",
     "moment": "*",
     "stripe": "*",
-    "lodash": "1.0.x",
     "coffee-script": "1.4.x",
-    "underscore": "*",
     "mongoskin": "*",
     "nconf": "*",
     "icalendar": "git://github.com/lefnire/node-icalendar#master",
@@ -26,7 +24,8 @@
     "resolve": "~0.2.3",
     "expect.js": "~0.2.0",
     "derby-i18n": "git://github.com/switz/derby-i18n#master",
-    "relative-date": "~1.1.1"
+    "relative-date": "~1.1.1",
+    "lodash": "~1.2.1"
   },
   "private": true,
   "subdomain": "habitrpg",
diff --git a/src/app/algos.coffee b/src/app/algos.coffee
deleted file mode 100644
index 1e9ab67cdf..0000000000
--- a/src/app/algos.coffee
+++ /dev/null
@@ -1,74 +0,0 @@
-XP = 15
-HP = 2
-
-priorityValue = module.exports.priorityValue = (priority='!') ->
-  switch priority
-    when '!' then 1
-    when '!!' then 1.5
-    when '!!!' then 2
-    else 1
-
-module.exports.tnl = (level) ->
-  if level >= 100
-    value = 0
-  else
-    value = Math.round(((Math.pow(level,2)*0.25)+(10 * level) + 139.75)/10)*10 # round to nearest 10
-  return value
-
-###
-  Calculates Exp modificaiton based on level and weapon strength
-  {value} task.value for exp gain
-  {weaponStrength) weapon strength 
-  {level} current user level
-  {priority} user-defined priority multiplier
-###
-module.exports.expModifier = (value, weaponStr, level, priority='!') ->
-  str = (level-1) / 2 # ultimately get this from user
-  totalStr = (str + weaponStr) / 100
-  strMod = 1 + totalStr
-  exp = value * XP * strMod * priorityValue(priority)
-  return Math.round(exp)
-
-###
-  Calculates HP modification based on level and armor defence
-  {value} task.value for hp loss
-  {armorDefense} defense from armor
-  {helmDefense} defense from helm 
-  {level} current user level
-  {priority} user-defined priority multiplier
-###
-module.exports.hpModifier = (value, armorDef, helmDef, shieldDef, level, priority='!') ->
-  def = (level-1) / 2 # ultimately get this from user?
-  totalDef = (def + armorDef + helmDef + shieldDef) / 100 #ultimate get this from user
-  defMod = 1 - totalDef
-  hp = value * HP * defMod * priorityValue(priority)
-  return Math.round(hp * 10)/10 # round to 1dp
-
-###
-  Future use
-  {priority} user-defined priority multiplier
-###
-module.exports.gpModifier = (value, modifier, priority='!', streak, model) ->
-  val = value * modifier * priorityValue(priority)
-  if streak and model
-    streakBonus = streak / 100 + 1 # eg, 1-day streak is 1.1, 2-day is 1.2, etc
-    afterStreak = val * streakBonus
-    model.set('_streakBonus', afterStreak - val) if (val > 0) # can we do this without model? just global emit?
-    return afterStreak
-  else
-    return val
-
-###
-  Calculates the next task.value based on direction
-  Uses a capped inverse log y=.95^x, y>= -5
-  {currentValue} the current value of the task
-  {direction} up or down
-###
-module.exports.taskDeltaFormula = (currentValue, direction) ->
-  if currentValue < -47.27 then currentValue = -47.27
-  else if currentValue > 21.27 then currentValue = 21.27
-  delta = Math.pow(0.9747,currentValue)
-  return delta if direction is 'up'
-  return -delta
-	
-	
diff --git a/src/app/character.coffee b/src/app/character.coffee
index 15b7563625..87159d2b56 100644
--- a/src/app/character.coffee
+++ b/src/app/character.coffee
@@ -1,6 +1,6 @@
 browser = require './browser'
 items = require './items'
-algos = require './algos'
+algos = require 'habitrpg-shared/script/algos'
 
 moment = require 'moment'
 _ = require 'underscore'
@@ -22,7 +22,6 @@ module.exports.app = (appExports, model) ->
     owned = user.get('items')
     # unless they're already at 0-everything
     if parseInt(owned.armor)>0 or parseInt(owned.head)>0 or parseInt(owned.shield)>0 or parseInt(owned.weapon)>0
-      console.log 'test'
       # find a random item to lose
       until loseThisItem
         #candidate = {0:'items.armor', 1:'items.head', 2:'items.shield', 3:'items.weapon', 4:'stats.gp'}[Math.random()*5|0]
diff --git a/src/app/debug.coffee b/src/app/debug.coffee
index a11541a5fc..32a1b82589 100644
--- a/src/app/debug.coffee
+++ b/src/app/debug.coffee
@@ -1,5 +1,5 @@
 moment = require 'moment'
-algos = require './algos'
+algos = require 'habitrpg-shared/script/algos'
 
 module.exports.app = (appExports, model) ->
   user = model.at('_user')
diff --git a/src/app/index.coffee b/src/app/index.coffee
index 6ff0c4e920..a3c1275b4a 100644
--- a/src/app/index.coffee
+++ b/src/app/index.coffee
@@ -117,7 +117,7 @@ ready (model) ->
   user = model.at('_user')
   model.setNull '_user.apiToken', derby.uuid()
 
-  require('./scoring').cron(model)
+  #FIXME require('habitrpg-shared/script/algos').cron(user)
 
   require('./character').app(exports, model)
   require('./tasks').app(exports, model)
diff --git a/src/app/tasks.coffee b/src/app/tasks.coffee
index 50d5c273c5..44a237dd20 100644
--- a/src/app/tasks.coffee
+++ b/src/app/tasks.coffee
@@ -1,12 +1,35 @@
-scoring = require './scoring'
+algos = require 'habitrpg-shared/script/algos'
 helpers = require 'habitrpg-shared/script/helpers'
-_ = require 'underscore'
+_ = require 'lodash'
 moment = require 'moment'
-character = require './character'
 
+
+###
+  Make scoring functionality available to the app
+###
 module.exports.app = (appExports, model) ->
+  character = require './character'
   user = model.at('_user')
 
+  ###
+    algos.score wrapper for habitrpg-helpers to work in Derby. We need to do model.set() instead of simply setting the
+    object properties, and it's very difficult to diff the two objects and find dot-separated paths to set. So we to first
+    clone our user object (if we don't do that, it screws with model.on() listeners, ping Tyler for an explaination),
+    perform the updates while tracking paths, then all the values at those paths
+  ###
+  score = (user, taskId, direction) ->
+    uObj = _.cloneDeep user.get() # need to clone, else derby won't catch model.set()'s after obj property sets
+    tObj = uObj.tasks[taskId]
+
+    # Stuff for undo
+    tObjBefore = _.cloneDeep tObj
+    tObjBefore.completed = !tObjBefore.completed if tObj.type in ['daily', 'todo']
+    setUndo uObj.stats, tObjBefore # set previous state for undo
+
+    paths = {}
+    algos.score(uObj, tObj, direction, {paths:paths})
+    _.each paths, (v,k) -> user.set(k,helpers.dotGet(k, uObj))
+
   appExports.addTask = (e, el) ->
     type = $(el).attr('data-task-type')
     newModel = model.at('_new' + type.charAt(0).toUpperCase() + type.slice(1))
@@ -28,26 +51,24 @@ module.exports.app = (appExports, model) ->
     model.unshift "_#{type}List", newTask
     newModel.set ''
 
-  appExports.del = (e, el) ->
+  appExports.del = (e) ->
     # Derby extends model.at to support creation from DOM nodes
     task = e.at()
     id = task.get('id')
 
     history = task.get('history')
-    if history and history.length>2
+    if history and history.length > 2
       # prevent delete-and-recreate hack on red tasks
       if task.get('value') < 0
-        result = confirm("Are you sure? Deleting this task will hurt you (to prevent deleting, then re-creating red tasks).")
-        if result != true
-          return # Cancel. Don't delete, don't hurt user
-        else
+        if confirm("Are you sure? Deleting this task will hurt you (to prevent deleting, then re-creating red tasks).") is true
           task.set('type','habit') # hack to make sure it hits HP, instead of performing "undo checkbox"
-          scoring.score(model, id, direction:'down')
+          score(user, id, 'down')
+        else
+          return # Cancel. Don't delete, don't hurt user
 
         # prevent accidently deleting long-standing tasks
       else
-        result = confirm("Are you sure you want to delete this task?")
-        return if result != true
+        return unless confirm("Are you sure you want to delete this task?") is true
 
     #TODO bug where I have to delete from _users.tasks AND _{type}List,
     # fix when query subscriptions implemented properly
@@ -113,14 +134,9 @@ module.exports.app = (appExports, model) ->
     Call scoring functions for habits & rewards (todos & dailies handled below)
   ###
   appExports.score = (e, el) ->
-    task= model.at $(el).parents('li')[0]
-    taskObj = task.get()
+    task = model.at $(el).parents('li')[0]
     direction = $(el).attr('data-direction')
-
-    # set previous state for undo
-    setUndo _.clone(user.get('stats')), _.clone(taskObj)
-
-    scoring.score(model, taskObj.id, direction)
+    score(user, task.get('id'), direction)
 
   ###
     This is how we handle appExports.score for todos & dailies. Due to Derby's special handling of `checked={:task.completd}`,
@@ -129,13 +145,7 @@ module.exports.app = (appExports, model) ->
   user.on 'set', 'tasks.*.completed', (i, completed, previous, isLocal, passed) ->
     return if passed? && passed.cron # Don't do this stuff on cron
     direction = if completed then 'up' else 'down'
-
-    # set previous state for undo
-    taskObj = _.clone user.get("tasks.#{i}")
-    taskObj.completed = previous
-    setUndo _.clone(user.get('stats')), taskObj
-
-    scoring.score(model, i, direction)
+    score(user, i, direction)
 
   ###
     Undo
diff --git a/src/app/viewHelpers.coffee b/src/app/viewHelpers.coffee
index c224dd308b..25a0d6af16 100644
--- a/src/app/viewHelpers.coffee
+++ b/src/app/viewHelpers.coffee
@@ -1,5 +1,5 @@
 _ = require 'underscore'
-algos = require './algos'
+algos = require 'habitrpg-shared/script/algos'
 items = require('habitrpg-shared/script/items').items
 helpers = require('habitrpg-shared/script/helpers')
 
diff --git a/src/server/api.coffee b/src/server/api.coffee
index b7b8af798b..1b96fd194c 100644
--- a/src/server/api.coffee
+++ b/src/server/api.coffee
@@ -1,9 +1,9 @@
 express = require 'express'
 router = new express.Router()
 
-scoring = require '../app/scoring'
 _ = require 'underscore'
-{ tnl } = require '../app/algos'
+algos = require 'habitrpg-shared/script/algos'
+{ tnl } = algos
 validator = require 'derby-auth/node_modules/validator'
 check = validator.check
 sanitize = validator.sanitize
@@ -255,7 +255,8 @@ scoreTask = (req, res, next) ->
     model.refList "_#{type}List", "_user.tasks", "_user.#{type}Ids"
     model.at("_#{type}List").push task
 
-  delta = scoring.score(model, taskId, direction)
+  #FIXME
+  delta = algos.score(user.get(), taskId, direction)
   result = model.get '_user.stats'
   result.delta = delta
   res.json result
diff --git a/src/server/deprecated.coffee b/src/server/deprecated.coffee
index 3fcef2c8a3..68d5d2030d 100644
--- a/src/server/deprecated.coffee
+++ b/src/server/deprecated.coffee
@@ -1,7 +1,6 @@
 express = require 'express'
 router = new express.Router()
 
-scoring = require '../app/scoring'
 _ = require 'underscore'
 icalendar = require('icalendar')
 api = require './api'