You're 15-10 this season. Your friend is 12-8. Who's the better player?

The records can't tell you. Your 15 wins might have come against beginners at open play, and their 12 might have come against the best players in the league. A win-loss record counts every opponent the same, so it says very little about skill.

An ELO rating weighs each result by who it came against. This guide covers how the formula works, how much one match moves the number, how UTR and NTRP translate into ELO points, and the exact rules RecAce uses when it calculates yours.

Quick Answers

What is an ELO rating in tennis or pickleball? A number that goes up when you win and down when you lose, where the size of the change depends on how strong your opponent was. Beating a stronger player earns a lot of points, and beating a weaker one earns only a few.

What is a good ELO rating? RecAce starts everyone at 1200, which lines up with roughly a UTR 6 or an NTRP 4.0 tennis player. Around 1500 is a strong club player (UTR 7.5, NTRP 4.5), and above 1800 is high-level competitive play (UTR 9, NTRP 5.0).

How many points does one match change my ELO? Between two evenly rated players, a win adds about 16 points and a loss takes away about 16. Beat someone 100 points above you during your first 10 matches and you gain about 26.

How many matches until my ELO is accurate? About 10. Your first 10 matches move the rating fastest, and after 30 it settles into smaller steps.

Is ELO the same as UTR or DUPR? No. UTR uses its own algorithm based on game margins, and DUPR uses its own model as well. ELO is older and simpler, and it only needs the people you play and who won.

How the ELO Formula Works

The physicist Arpad Elo built the system for chess in the 1960s. (Elo is his last name, which is why chess writers spell it with one capital letter. Most apps, including RecAce, write it ELO.) It runs in two steps after every match.

Step 1: predict the result. The rating gap between you and your opponent turns into an expected score between 0 and 1:

Expected score = 1 / (1 + 10^((opponent rating − your rating) / 400))

If you're both rated 1200, your expected score is 0.50, a coin flip. If they're 100 points higher, it drops to 0.36.

Step 2: compare the prediction to what happened. A win scores 1 and a loss scores 0. Your rating changes by:

Change = K × (actual score − expected score)

K is the "speed" setting, which caps how far one match can move you. When the result matches the prediction, you barely move. When it doesn't, you move a lot. That's the whole system. A favorite who wins as expected gains little, and an underdog who wins gets the big jump.

Worked Examples: How Much One Match Moves You

These use the same numbers RecAce uses, with a K of 40 during your first 10 matches and 32 for matches 11 through 30.

SituationExpected scoreWinLoss
Match 5: you 1200 vs. opponent 13000.36+26−14
Match 5: you 1300 vs. opponent 12000.64+14−26
Match 20: you 1200 vs. opponent 12000.50+16−16
Match 20: you 1200 vs. opponent 14500.19+26−6

The last row shows why a run of losses to stronger players doesn't sink you. Losing to someone 250 points above you costs about 6 points, because the formula expected it. Beating them pays more than four times that.

ELO Rating Differences and Win Probability

Since the expected score is a win probability, the rating gap alone tells you how often the stronger player should win.

ELO differenceHigher-rated player's win probability
0 points50%
50 points57%
100 points64%
150 points70%
200 points76%
300 points85%
400 points91%

A 100-point favorite still loses about one match in three, so a single result never settles who's better. This table also explains losing streaks against higher-rated opponents. If you've been playing people 150 points above you, losing most of those matches is exactly what the math predicts, and your rating barely moves.

UTR and NTRP to ELO Conversion Chart

ELO numbers mean nothing until you tie them to a scale you already know. RecAce converts UTR at a flat 200 ELO points per UTR point, so UTR 6.0 equals 1200. With that spacing, a 1.0 UTR gap predicts the better player wins about 76% of the time.

NTRP converts on a curve, because the jump from 4.5 to 5.0 is a much bigger step in real skill than the jump from 3.0 to 3.5. RecAce fits the curve to known NTRP and UTR equivalents so the two scales land in the same place.

NTRPELOUTRELO
2.54913.0600
3.07044.0800
3.59395.01000
4.011956.01200
4.514737.01400
5.017728.01600
5.520939.01800

Pickleball starting levels take the same UTR or NTRP entries, set separately from tennis. Whichever scale you pick, use it for your opponents too so everyone lands on the same footing. If you only know one tennis rating, our UTR to NTRP converter translates between the two.

Start your rating from the level you already have

Set your UTR or NTRP as a starting level and log your next match. Your ELO picks up from there instead of a generic 1200.

Find My ELO Rating

How RecAce Calculates Your ELO

A plain ELO calculator has two inputs: the two ratings. RecAce adds context from the match you logged. Here's every rule it applies.

Your starting level

Everyone starts at 1200 unless you set a starting level. RecAce asks for it when you sign up, and you can change it any time in Profile > Edit profile. Enter the UTR or NTRP you had when you began logging matches, and RecAce converts it with the chart above. If you were a 3.0 when you started and you're a 4.0 now, your graph shows that climb instead of a strange drop from 1200 in the first few weeks. Tennis and pickleball each get their own starting level.

K changes as you play more

Opponents with a known rating start at that rating

When you record an opponent's NTRP or UTR, they enter the math at that level instead of the 1200 default. That stops a strong player you've only faced once from counting as an average one. RecAce also keeps pulling that opponent back toward their recorded rating, because a USTA rating built on dozens of matches says more than the few matches they've played against you. That pull weakens as your head-to-head history with them grows.

How close the match was

RecAce scores every match for match quality, so a 7-6 7-6 battle and a 6-1 6-0 rout don't count the same. How that changes K depends on the rating gap:

The effect fades smoothly. At a gap of about 100 points the score doesn't change K at all, and by 200 points the effect has fully flipped. In the last example from the table above, beating a player 250 points above you in a tight match pays about 32 points instead of 26.

Doubles uses your team's rating

In doubles, RecAce averages the two opponents' ratings and compares them with your team. If your partner has a recorded rating, your team enters the match at the average of you and your partner. If not, the team counts at your rating. Your main rating takes singles and doubles together; Pro can also show a singles-only and a doubles-only rating, each a separate calculation. Tennis and pickleball are always separate. Being a 4.0 in tennis doesn't make you a 4.0 in pickleball.

Practice doesn't count

Matches you tag as practice don't count, so a hitting session where nobody kept score can't drag your rating down.

RecAce rating chart: a tennis rating over a year of matches, from the starting level up to a best of 1,336

Reading Your ELO Trend

Your rating is free for every account. The dashboard shows your current number with context around it:

With Pro, the rating card in Analytics adds the full history chart: every match as a point you can tap, your best marked, and All, Singles, and Doubles views.

The shape of the line tells you more than any single point on it:

Opponent Difficulty: ELO Applied

The rating gap turns into a predicted win probability for every opponent in your history. RecAce shows it as your odds (a Pro feature, and a core part of RecAce's opponent scouting). A simple way to read it:

This changes how you read a month. Go 3-7 against ten tough opponents and your rating can still rise. Go 8-2 against players you're favored to beat and it barely moves. If you want to improve, look for even and tough matchups. If you want confidence before a tournament, a few easier sessions are fine. The mix of difficulty is one of the five stats every player should track.

Ratings also drive the win predictions in head-to-head and scouting views. Those start from the rating gap and then blend in your actual record against that opponent, so after several meetings your real history carries more of the weight.

Why Pickleball Players Use ELO

DUPR has become the standard pickleball rating, but it only sees matches that reach DUPR through events, clubs, or scores your opponents validate. Most open play never gets entered. Self-assessed levels (3.0, 3.5, 4.0) are even less reliable, since nearly everyone rates themselves half a level high.

A personal ELO fills that gap. It updates from every match you log, including Tuesday night open play, and it measures you against the people you actually play. Use it alongside DUPR, not instead of it. If your ELO is climbing and your DUPR isn't, you may just not have many recorded DUPR matches yet.

Why Tennis Players Use ELO

UTR and NTRP only update from matches they can see, which means sanctioned events and leagues. Hitting sessions, club ladders, and friendly sets never reach them. A personal ELO counts all of it:

If you're heading into your first league season, your ELO is a useful second opinion on which NTRP level to enter.

Common ELO Misconceptions

"My ELO dropped, so I got worse"

Not necessarily. Tougher opponents mean more expected losses, and the rating dips while it adjusts. Check who you played. If your graph opens with a big early slide, set your starting level in Profile > Edit profile so the calculation stops correcting from 1200.

"I beat someone with a higher ELO, so the rating is wrong"

ELO predicts probability, not certainty. A player 100 points below wins about 36% of the time. One upset is normal. Ten in a row means the ratings will move to match.

"I need hundreds of matches for a real rating"

You don't. After 10 matches the rating is out of its provisional stage, and by 30 it's stable enough that a single match moves it only a little.

"ELO ignores doubles"

Doubles count, with your partner factored into the team rating. Your main rating processes both formats in date order. Pro's Singles and Doubles views calculate each format on its own, so neither will equal the combined number. That's expected.

Start Tracking Your Rating

Your win-loss record gives you the count. Your ELO accounts for who you played.

Logging matches in RecAce's free match stat tracker is free, and so is your rating: it updates after every match on every account. Pro adds the rating history chart, singles and doubles views, and your odds against each opponent, and logging 5 matches earns you 30 days of Pro free. Log your tennis and pickleball matches, set your Starting Level, and the rating builds from there. If you upgrade later, the chart covers your whole history, not just matches after the upgrade. The ELO ratings guide covers the same rules in reference form.

For the other rating systems you'll run into, read our guides to NTRP ratings, UTR vs. NTRP, and the UTR to NTRP conversion chart.

See Your ELO Rating Free →