i went ahead and re-ran the #'s based on what you have listed, subbing in game #1 #'s from page 2.
and i thought i'd give you another opinion.
| #1 | #2 | #3 |
| average |
| real |
| diff |
score | 27 | 15 | 29 |
| 23.67 |
| 24 |
| 0.33 |
tot yds | 281 | 321 | 394 |
| 332.00 |
| 391 |
| 59.00 |
pass | 31 | 48 | 33 |
| 37.33 |
| 32 |
| -5.33 |
comp % | 51 | 40 | 54 |
| 48.33 |
| 58.1 |
| 9.77 |
pass yds | 194 | 197 | 225 |
| 205.33 |
| 212 |
| 6.67 |
int | 1 | 2 | 0 |
| 1.00 |
| 1.4 |
| 0.40 |
drop | 3 | 3 | 1 |
| 2.33 |
| 4.5 |
| 2.17 |
sack | 6 | 4 | 2 |
| 4.00 |
| 3 |
| -1.00 |
rush | 44 | 38 | 49 |
| 43.67 |
| 38 |
| -5.67 |
rush yds | 87 | 124 | 169 |
| 126.67 |
| 165 |
| 38.33 |
yd/c | 2 | 3.4 | 3.4 |
| 2.93 |
| 4.3 |
| 1.37 |
yd/c #1 rb | 3.9 | 4.1 | 3.9 |
| 3.97 |
| 5.1 |
| 1.13 |
tfl | 6 | 7 | 7 |
| 6.67 |
| 5.5 |
| -1.17 |
fumble | 1 | 0 | 0 |
| 0.33 |
| 1.1 |
| 0.77 |
punt | 44.9 | 42.1 | 46.4 |
| 44.47 |
| 38.5 |
| -5.97 |
penalty | 6 | 6 | 6 |
| 6.00 |
| 5.7 |
| -0.30 |
pen yds | 35 | 50 | 45 |
| 43.33 |
| 52 |
| 8.67 |
hold | 3 | 4 | 4 |
| 3.67 |
| 0 |
| -3.67 |
facemask | 1 | 0 | 0 |
| 0.33 |
| 0 |
| -0.33 |
false start | 0 | 1 | 0 |
| 0.33 |
| 0 |
| -0.33 |
pass int | 0 | 0 | 0 |
| 0.00 |
| 0 |
| 0.00 |
delay | 2 | 1 | 0 |
| 1.00 |
| 0 |
| -1.00 |
my observations:
-score is right one, obviously
-total yds is a little low, but is, imo, negligible having only 3 tests so far.
-pass diff is negligible considering sample size. depending the way the games played out, the difference is nothing unusual.
-comp % is low, but could be the result of simply not being used to new sliders yet. (as you mentioned)
-pass yds is very close
-int is close
-drops are low, especially when considering the comp % is low (would expect them to be conversely related) meaning the incomp are due to the qb inaccuracy. raising qb accuracy might lower drops even more. imo, qb accuracy needs to raise (might be simple as the player getting used to the settings) and wr catching needs to be lowered some.
-sacks are not off on average, but the variance is too large (too big of a range). the small sample size makes it hard to determine anything.
-rushes are little high, like passes. also like passes, this could simply be a result of the games. however, these should also be conversely related. (not exactly, but over time) the more you pass, the less you run. but, again, sample size makes this too hard to determine any trend. simple solution might be too many plays per game.
-rush yds are low. i'll divert to oraeon's description on this, as he actually played the games. outside rushes were the problem according to him.
-y/c and y/c main rb are both low, and again, i'll leave it too oraeon to explain that.
-tfl is a little high, which is probably closely related to the low outside rush yds, per oraeon's description.
-fumble is little low, but considering sample size, not unusual.
-punt yds is long by about 5 yds. should lower punt power.
-penalty #'s are very close.
-penalty yds are a little low, but that can be a result of the types of penalties.
overall, looks pretty accurate, but one big thing is that sample size is simply too small to make any real educated adjustments. too small to see any trends.
also, as oraeon said, individual players will obviously be different skills and will need to tweak to likeness. but this is definitely a solid starting point.