When OBP and BB/100 don’t make sense relative to each other which should be trusted (ie which does the engine use)?

If you're looking at how often a player will walk then definitely BB/100. OBP is influenced not just by walks, but also hit by pitches and sac flies.

If the # stats are Log 5 normalized with historical average stats, shouldn't they change every time a new season's worth of data is created and added to the historical average?

sounds logical to me.

Posted by StatFiend on 10/13/2026 3:36:00 PM (view original):
When OBP and BB/100 don’t make sense relative to each other which should be trusted (ie which does the engine use)?

The metric used by the sim engine is BB/(PA-HBP), so BB/100PA is the better stat to use when drafting

Posted by _JJP_ on 10/13/2026 6:37:00 PM (view original):
If the # stats are Log 5 normalized with historical average stats, shouldn't they change every time a new season's worth of data is created and added to the historical average?

# stats aren't used by the sim engine to determine the outcome of a PA, so it really doesn't matter.

Posted by skunk206 on 10/13/2026 7:49:00 PM (view original):

Posted by _JJP_ on 10/13/2026 6:37:00 PM (view original):
If the # stats are Log 5 normalized with historical average stats, shouldn't they change every time a new season's worth of data is created and added to the historical average?

# stats aren't used by the sim engine to determine the outcome of a PA, so it really doesn't matter.

What does the "Every step...." note at the top mean then?



https://image.slideserve.com/41801/pa-decision-tree-normalization-n.jpg

it uses the league average from that individual season. so a guy who hit .300 when the league average was .250 that season has a better chance of getting a hit then if the league average was .280 that season. the only input that matters is that individual league and season average. the historical total league average is irrelevant to the calculation.

OK, so is what you're saying that the + stats get used to determine the outcome of a PA? That seems to match what the first bold part below says:

The other biggest item to note when drafting is normalization. WhatIfSports.com uses log5 normalization in which the batter and pitcher's statistics are compared to the league averages from their respective seasons. Statistical categories indicated by the # represent how this player would fare against the historically average player. In this case, players who were great hitters in poor offensive years like 1968, get a little more credit, while those with similar raw numbers from a great offensive season like 1930 will not fare as well. Log5 (#) values are the single most important representations of a player's value in SimLeague Baseball and are thus highly incorporated in salaries. (https://www.whatifsports.com/mlb-l/PlayersGuide.asp?t=5)

But the second bolded part brings me back to my first question. If the # stats are not used in determining the outcome of a PA, but they are the single most important representation of a player's value (and tied to salary), shouldn't that number change year to year just as the "historically average player" changes?

I would think so. But I stink at this.

Agreed that in theory they should. If I had to guess, I'd say that they're using the numbers from the league averages table (https://www.whatifsports.com/mlb-l/league_averages.asp?Action=view), which stops at 2011, for the # stats. Even if they updated them every year, though, players would all move in the same direction relative to each other, so since those numbers are not used in the decision tree I guess they figure it's not worth the hassle.

Agreed that this doesn't make a difference to individual PAs since # stats are not used in the decision tree.

But if the "historically average player" that's used to create the # stats is the one at the bottom row of that chart (just below 2011), then isn't every player added from 2012-2026 using a separate standard for how # stats are determined? The last 15 years of data won't be factored into the historically average player data and each year it would get progressively further from the 2011 baseline .

It won't effect an individual plate appearance, but doesn't it have a big effect on salaries becuase the # stats determine the salary? New players have a salary based on a 15 year old formula that doesn't account for the recent changes in the game and current players have a salary that remains unchanged from what the average was 15 years ago (with a blip a few years after that for the usage-based increases/decreases).

Dynamic salaries were a mess and still are when comparing newer-season player values to older seasons. The salary formula will need an update, and salaries reset across the board at some point.

Leading 5-0 in the 7th-9th the game chose to pitch both of my setup As instead of my Long A, Long B or Setup B… Why? How do I set things up so it doesn’t waste my best pitchers when any pitcher will do?

the most effective way to ensure your best relievers are only used in tight situations is to have them as Closer A so they will only come in when the game is tied or you're slightly ahead. this of course has drawbacks though since they won't come in when you're losing by one run.

anything you do has benefits and drawbacks though, you can't get the sim to do exactly what you'd do in every situation.