nfldb & nflgame data mismatch

[ochawkeye]

I'm finding some instances where nfldb data does not match what I would expect based on last years nflgame data. For example:

from __future__ import print_function
y, w = 2013, 7

import nflgame
games = nflgame.games(year=y, week=w, kind='REG')
players = nflgame.combine_game_stats(games)
for x in players:
    if x.name == 'R.Gould' or x.name == 'G.Gano':
        print(x.name, getattr(x, 'kicking_xpmade'))

print('-'*79)

import nfldb
db = nfldb.connect()
q = nfldb.Query(db)
q.game(season_year=y, season_type='Regular', week=w)
q.player(full_name='Graham Gano')
for pp in q.as_aggregate():
    print(pp.player.gsis_name, getattr(pp, 'kicking_xpmade'))

q = nfldb.Query(db)
q.game(season_year=y, season_type='Regular', week=w)
q.player(full_name='Robbie Gould')
for pp in q.as_aggregate():
    print(pp.player.gsis_name, getattr(pp, 'kicking_xpmade'))

Produces the following:

G.Gano 3
R.Gould 5
-------------------------------------------------------------------------------
G.Gano 2
R.Gould 4

Where are the missing extra points? Am I simply using incorrect criteria in nfldb queries?

[BurntSushi]

nfldb is play-by-play data only. Your nflgame code is using combine_game_stats. Instead, could you try it with combine_play_stats? Hopefully this will produce equivalent results.

Regrettably, extra points are, by far, the statistic with the most volatility.

[ochawkeye]

Indeed, changing to combine_play_stats gives identical results. I'm now reminded of a long ago issue reporting against nflgame on this very topic.

[ochawkeye]

No strong desire to hunt down each and every discrepancy, but these are a few I stumbled across when trying to reconcile scores for week 1.

Blair Walsh (MIN, K): 4 XP made vs. 3 XP in nfldb
Cody Parkey (PHI, K): 4 XP made vs. 3 XP in nfldb
Dexter McCluster (TEN, RB): -2 punt return yards <-- (this one actually looks correct, nflgame maxed this one incorrectly)