How Fantasy Rankings Are Built and How to Adjust Them to Your League
How fantasy baseball rankings turn projected stats into one number using z-scores and standings gain points, with worked examples with made-up players, and how to adjust a list to your league format.
Every fantasy baseball ranking you will see this winter, whether it comes from a website, a magazine or a spreadsheet a friend sent around, is trying to answer the same question: how much is one player's projected season worth to a team in a league like yours? The answer is a number, and the order of those numbers is the ranking. The interesting part is how the number is made, because the method decides who ends up at the top and where your league's rules should move them.
There are two main methods. The first turns each statistic into a z-score, which measures how far a player is above or below the average of the draftable pool. The second, standings gain points, measures how many places in the standings a player's statistics would move a team. Both start from projections, both subtract a replacement level, and both can be adapted. They give similar results for most players and different results for a few, and the differences are where good drafters find value.
This article builds a ranking from scratch with invented players, so that every number can be checked by hand. None of the players below is real, and none of the numbers is a forecast. At the end it covers the adjustments that matter most: league size, scoring format, positions, innings limits and the difference between a snake draft and an auction.
What a ranking is actually measuring
A ranking combines four inputs. The first is a projection: expected totals for each statistic for the coming season, usually produced by averaging several projection systems or by a single analyst. The second is a scoring format, because a home run is worth something different in a league that counts it than in one that does not. The third is the size of the player pool that will actually be drafted. The fourth is a replacement level, the quality of the player you could pick up for free during the season.
The format most rankings assume is rotisserie, or "roto", with five hitting categories and five pitching categories. The hitting categories are batting average, home runs, runs, runs batted in and stolen bases. The pitching categories are wins, saves, strikeouts, earned run average and WHIP, which is walks plus hits per inning pitched. Each team is ranked in each category and earns points for its rank, so in a 12-team league the best team in a category earns 12 points and the worst earns 1. The team with the most total points wins. Head-to-head categories leagues use the same ten statistics but score them week by week, and points leagues assign a fixed value to each event.
The projections themselves are the least certain input. A projection is an average over a range of outcomes, and a ranking built on it will be wrong about many individual players. The ranking is a tool for sorting, not a prediction.
Method one: the z-score
The z-score is the standard measure from statistics. For any category, take a player's projected total, subtract the average total of the draftable pool, and divide by the spread of that pool, which is its standard deviation. The result says how many standard deviations above or below average the player is. A z-score of +2 is far above average in that category, 0 is average, and -1 is below.
Here is a worked example. Suppose the pool of draftable hitters in a 12-team league averages 20 home runs with a standard deviation of 8, 75 runs with a standard deviation of 15, 72 RBI with a standard deviation of 16, and 8 stolen bases with a standard deviation of 8. Two invented hitters, Player A and Player B, are projected as follows.
| Category | Pool average | Pool spread | Player A | Player B |
|---|---|---|---|---|
| Home runs | 20 | 8 | 35 | 12 |
| Runs | 75 | 15 | 95 | 85 |
| RBI | 72 | 16 | 100 | 55 |
| Stolen bases | 8 | 8 | 4 | 38 |
Player A's z-scores are (35 - 20) / 8 = +1.88 for home runs, (95 - 75) / 15 = +1.33 for runs, (100 - 72) / 16 = +1.75 for RBI, and (4 - 8) / 8 = -0.50 for stolen bases. Player B's are -1.00 for home runs, +0.67 for runs, -1.06 for RBI and +3.75 for stolen bases. Adding the four, Player A has +4.46 and Player B has +2.35.
Batting average needs special handling, and it is where many amateur spreadsheets go wrong. A .300 hitter with 300 at-bats helps a team's average less than a .280 hitter with 650. The right measure is the number of hits above what an average hitter would have collected in the same at-bats. If the pool hits .260, Player A, at .275 in 560 at-bats, has 154 hits against an expected 145.6, which is 8.4 hits above average. Player B, at .270 in 600 at-bats, has 162 hits against an expected 156, which is 6.0 above. If the spread of that measure across the pool is 7 hits, their z-scores are +1.20 and +0.86. The totals become +5.66 for A and +3.21 for B.
The z-score method rests on one assumption: that a standard deviation in one category is worth the same as a standard deviation in another. That is not exactly true, and it is the weakness the second method is designed to fix.
Method two: standings gain points
In a roto league, what counts is not how far above average you are but how many places you climb. Standings gain points, or SGP, convert each statistic into the number of standings places it is worth in a league like yours.
The key quantity is the gap: the average difference in a category between one team and the next in the standings. If teams are separated by 8 home runs on average from rank to rank, then 8 home runs is worth one standings point, and a player with 35 home runs contributes 35 / 8 = 4.4 points from that category. The gaps come from past standings in a league of the same size and format, or from published values, and they differ by category.
Take invented gaps for a 12-team league: 8 home runs, 22 runs, 22 RBI, 9 stolen bases, and .0020 in team batting average. Player A's counting statistics are worth 35 / 8 + 95 / 22 + 100 / 22 + 4 / 9 = 4.38 + 4.32 + 4.55 + 0.44 = 13.68 points. Player B's are worth 12 / 8 + 85 / 22 + 55 / 22 + 38 / 9 = 1.50 + 3.86 + 2.50 + 4.22 = 12.09 points.
For batting average, divide the hits above average by the team's total at-bats to get the change in team average, then divide by the gap. If a team has about 5,600 at-bats, Player A's 8.4 extra hits add .0015 to the team average, which is 0.75 points, and Player B's 6.0 extra hits add .0011, or 0.54 points. The full totals are 14.43 for A and 12.63 for B, a gap of 1.8 points, which compares to the 2.45 gap from the z-score method. The two methods agree on the order and differ on the margin.
| Method | Player A | Player B | Gap |
|---|---|---|---|
| Sum of z-scores | 5.66 | 3.21 | 2.45 |
| Standings gain points | 14.43 | 12.63 | 1.80 |
The SGP method values the stolen-base specialist a little more than the z-score does here, because 38 stolen bases is a large number of places in a category where the gaps are small. In practice, which method you prefer matters less than whether your format and gaps are reasonable.
Replacement level and position
Both totals above are raw. A ranking needs to subtract what a team could get for free. That is the replacement level: the value of the best player available at no cost once the draft is over. Players are not worth their total; they are worth their total minus the replacement level at their position.
This is where position matters. In a typical league each team starts one catcher, and the 12th-best catcher is far worse than the 12th-best first baseman. Suppose that after summing z-scores, the last starting-quality catcher in the pool sits at -1.0 and the last starting-quality first baseman sits at +0.8. A catcher with a raw total of +2.0 then has a value over replacement of 3.0, and a first baseman with +2.6 has 1.8. The catcher ranks higher despite the lower raw total.
| Player | Raw total | Replacement at position | Value over replacement |
|---|---|---|---|
| Catcher (invented) | +2.0 | -1.0 | 3.0 |
| First baseman (invented) | +2.6 | +0.8 | 1.8 |
Replacement level also depends on the number of roster spots. Add a second catcher slot, as many leagues do, and the replacement level for catchers drops sharply, which raises the value of the good ones. Add a utility spot, and replacement levels for the other positions rise because the best leftover hitters now have somewhere to play.
Pitchers: volume matters
Pitchers add two wrinkles. First, ERA and WHIP, like batting average, are rates, so a pitcher's contribution depends on how many innings he throws. The measure is runs prevented, or baserunners prevented, relative to the pool average. Second, saves are scarce and heavily concentrated in a few relievers.
Take the invented starter and reliever below, with the pool average ERA at 4.00 and the pool WHIP at 1.28.
| Pitcher | Innings | ERA | WHIP | Runs saved vs pool | Baserunners saved vs pool |
|---|---|---|---|---|---|
| Starter | 180 | 3.50 | 1.15 | 10.0 | 23.4 |
| Reliever | 65 | 2.50 | 1.00 | 10.8 | 18.2 |
Runs saved is (4.00 - ERA) times innings, divided by 9. For the starter, that is 0.50 x 180 / 9 = 10.0. For the reliever, 1.50 x 65 / 9 = 10.8. Baserunners saved is (1.28 - WHIP) times innings: 0.13 x 180 = 23.4 and 0.28 x 65 = 18.2. The reliever's ERA looks far better, but over 65 innings he saves almost exactly as many runs as the starter does over 180. The starter then adds more strikeouts and wins, while the reliever adds saves, if he has the job. That is why rankings place a closer with a secure role above a setup man with identical ratios, and why the ratios of middle relievers are worth less than they look. The strategy side of that is in our bullpen strategy article.
Turning value into a ranking or a price
Once every player has a value over replacement, a snake-draft ranking simply sorts them from highest to lowest. An auction needs a price. Divide the league's total budget, split between hitters and pitchers in some proportion, by the sum of positive values over replacement, and multiply each player's value by the result.
For example, in a 12-team league with a $260 budget per team, the total pool of money is $3,120. If 68 percent goes to hitters, that is about $2,120. If the 168 hitters who are drafted have a combined value over replacement of 480 units, each unit is worth about $4.42. A hitter at 6.0 units is then worth about $26.50. The process for pitchers is the same, with their own budget share. More on how to use these prices is in our auction draft article, and on how to use a ranking in a snake draft in our snake draft article.
Adjusting a ranking to your league
A published ranking is built for a default league, usually 12 teams, 5x5 roto, standard positions. If yours is different, these are the changes that matter most, roughly in order of impact.
- League size. In a 15-team league, the pool is deeper and the replacement level lower, so every drafted player is worth more over replacement and depth matters. In a 10-team league the opposite is true, and waiver-wire pickups are good enough to cover most holes. See our deep league article for AL-only, NL-only and 15-team formats.
- Scoring categories. If your league uses on-base percentage instead of batting average, walks become valuable and the ranking should be rebuilt. If it adds holds or quality starts, relievers and innings eaters rise.
- Points leagues. A points league gives each event a fixed value, so no z-scores are needed: multiply each projected stat by its point value and add them up. The adjustments that remain are positional scarcity and replacement level.
- Roster rules. More bench spots lower the replacement level. Two catchers, a corner and middle infield slot, and a utility player each change the scarcity of positions.
- Innings limits. Many leagues cap total innings. Under a cap, every inning must be used well, so quality per inning, meaning ratios and strikeouts per inning, matters more than raw volume.
- Keepers. A keeper league changes the pool, since players kept off the board are no longer available. Ranking should be done on the pool that will actually be drafted.
- Platoon and two-way players. These cases break the simple model, since playing time is uncertain. See our platoon and two-way player article.
Common mistakes
The most common error is ranking by raw totals, adding up home runs, runs and steals as if a steal were worth the same as a home run. Each category has to be converted to a common scale first, which is what z-scores and standings gain points do. The second is ignoring innings: two pitchers with the same ERA are not the same value if one throws 200 innings and the other 120. The third is treating the ranking as a hard order. A ranking is built on projections whose error is large compared to the differences between neighbours, so players within a few places of each other are interchangeable. Many drafters prefer tiers, groups of players of roughly equal value, and use the ranking only to find where one tier ends and another begins.
A fourth mistake is trusting the same list in March and in June. Replacement level changes during the season as players are drafted, hurt and dropped, and a list built before the draft does not know what is on your waiver wire.
Common questions
Which is better, z-scores or SGP? For most leagues, they give similar orderings. SGP is more accurate for roto leagues if you have good gap data for your league; z-scores are easier and work for categories leagues of any kind.
Why do closers rank so high? Saves are scarce and concentrated. A closer's ratios help as much per inning as anyone's, but he throws far fewer innings, so most of his edge comes from saves and a secure role.
Should I draft strictly by the ranking? It is a good default. Departing from it should be a decision about your league or your roster, not a hunch.
Are the projections reliable? They are averages, and on any one player they can be far off. Treat them as the best available starting point.
Do these methods apply to money leagues? The methods are identical, but if you play for money, only do so where it is legal, play within your means, and if gambling is a problem for you or someone you know, the national helpline is 1-800-GAMBLER.
Where to go next
Once you have a ranking, the snake draft guide and the auction guide explain how to use it. For tougher formats, read the expert and deep league guide, and for what the real players did last season, use the leaders pages and the fantasy section of this site. The fantasy baseball topic page collects news that moves rankings.