Welcome back fam — July’s in the books, and we’re doing the whole month in one go this time. SaltyTrack logged 12,652 matches and $81.8 billion in Salty Bucks across July 1–31, and our model finished at 75.0% against the crowd’s 69.0%.[1] That’s a 6.1-point edge — 768 more correct calls than the crowd managed over the same board.
Beautiful family, welcome back! July was the richest month we’ve tracked all year — more money than May ($77.8B) and June ($79.3B), on fewer matches than May. It also produced the single most improbable result we’ve seen in a while: on the last night of the month, a fighter with 1/37th of the money on it walked out a winner. Grab a snack — here’s the full month: the numbers, who’s hot, the biggest upsets, the whale pots, the streaks, how our predictions held up, and a genuinely surprising thing we found about P-tier.
The thirty-one-day flyover of SaltyBet:
| Metric | Value |
|---|---|
| Matches | 12,652 |
| Total Wagered | $81.8 billion |
| Unique Fighters | 7,543 |
| Avg Matches/Day | ~408 |
| Avg Pot | $6.47M |
| Mode Split | 86.7% matchmaking / 13.3% tournament |
| Side Balance | Red 49.9% / Blue 50.1% |
7,543 unique fighters across 12,652 matches. Since every match needs two of them, that’s 25,304 fighter-slots filled by 7,543 distinct characters — the average fighter appeared about three times all month. That’s SaltyBet’s roster doing what it always does: an ocean of MUGEN characters, most of whom you’ll see once or twice before the matchmaker spins them back into the pool. New to how that churn works? The Beginner’s Guide covers the basics.
And the coin flip landed at 50.1% blue / 49.9% red — a 28-match difference across twelve and a half thousand matches. We say this every single report and we’ll say it again: the red/blue split is noise. It has always been noise. It will always be noise.
Here’s how the action sorted across the tiers:
| Tier | Matches | % of Total | Avg Pot |
|---|---|---|---|
| A | 5,377 | 42.5% | $5.73M |
| S | 4,042 | 31.9% | $7.04M |
| B | 2,659 | 21.0% | $6.55M |
| X | 317 | 2.5% | $7.17M |
| P | 257 | 2.0% | $11.28M |
A-tier and S-tier together were 74.4% of all the action — that’s where the matchmaking lives. X and P stayed rare as ever, together barely 4.5% of the card. And look at that right-hand column: P-tier drew the fattest average pots at $11.28M — nearly double A-tier’s, on one-twentieth the matches. That paradox has shown up in every report we’ve written. This month we finally dug into why, and the answer surprised us. (Scroll to The Weird Stuff.)
These fighters cleared the entire month without dropping a single match (6+ wins):
| Fighter | Record | Tier |
|---|---|---|
| Master marco | 10-0 | S |
| Guts | 9-0 | S |
| Lt.kurosawa EX | 9-0 | A |
| Nine the phantom | 9-0 | S |
| Dusk komachi | 8-0 | S |
| Flawless specter | 8-0 | S |
| Fw_gigan | 8-0 | A |
| Sol(ac) | 8-0 | S |
Master marco went 10-0 — the only fighter to reach double digits unbeaten across the whole month, and the cleanest run on the board. Six of the eight perfect records came out of S-tier, which tracks: S is where the genuinely strong characters live, and a strong character that keeps drawing S-tier opponents can run the table for weeks. Lt.kurosawa EX and Fw_gigan did it from A-tier instead, which is the harder version of the trick — A-tier is the deepest, busiest pool on the site (5,377 matches), so nine straight wins there means nine straight favorable rolls of the matchmaker.
And then there’s the other end. These fighters went the entire month without a single win (6+ losses):
| Fighter | Record | Tier |
|---|---|---|
| Joe higashi EX3 | 0-8 | S |
| Kaede kasiwagi | 0-8 | S |
| Masaki anzu | 0-8 | A |
| Anon yukinomiya | 0-7 | A |
| Asura | 0-7 | S |
| Battler ushiromiya | 0-7 | S |
Three fighters went 0-8, and five of the six worst records came from S and A — the tiers with the most matches, so the most chances to compile a bad month. Being winless in S-tier isn’t a statement about being the worst character on SaltyBet; it’s a statement about being the weakest character in a strong room. Tier placement is relative, which is exactly what our tier list explainer gets into.
Ranked by bet ratio — how lopsided the money was against the eventual winner:
| Date | Winner | Loser | Ratio | Tier |
|---|---|---|---|---|
| Jul 31 | Gomora ug | Kimexemaru | 37.5x | S |
| Jul 5 | Kitana | Bomber man | 24.1x | B |
| Jul 18 | Kyoushirou senryou | Dhalsim vp | 22.4x | A |
| Jul 2 | Maverick hunter x | Shiteyannyo | 20.1x | S |
| Jul 9 | Custom prier | Terra branford | 19.2x | B |
On the last night of the month, Gomora ug beat Kimexemaru at a 37.5-to-1 ratio — the most lopsided result of July and the biggest single-match shock we’ve logged since the spring. The crowd had made up its mind completely; roughly 97% of the pot sat on Kimexemaru. It lost anyway.
The more painful one was Kitana over Bomber man on July 5 at 24.1-to-1 — inside a $54.1 million pot. That’s the difference between an upset and a meltdown: Gomora ug’s shock happened in a quiet $1.6M pot, so almost nobody got hurt. Kitana’s happened while $52 million of crowd money was stacked on the other side. Same species of surprise, wildly different body count. If you’ve ever wondered why we keep telling people to size their bets to their confidence rather than their certainty — this is why. Our piece on the psychology of betting on AI fights digs into why crowds pile onto doomed favorites like this.
| Date | Matchup | Winner | Pot | Tier |
|---|---|---|---|---|
| Jul 27 | Vicviper vs Green lantern gg | Green lantern gg | $372.1M | B |
| Jul 24 | Daisy Princess vs Akito(coh) | Daisy Princess | $364.6M | P |
| Jul 1 | Cci kim vs Decade as kuuga | Cci kim | $331.9M | B |
| Jul 14 | Hakolen-sbs vs Galvatron | Galvatron | $291.2M | P |
| Jul 5 | Mukuro vs Ash crimson | Ash crimson | $263.7M | B |
The fattest pot of the month was $372.1 million on July 27, and the favorite held. In fact every one of the top five pots was won by the side the money was on — a clean sweep for the favorites at the high end. That’s the reassuring version of whale season: when the giant bets landed in July, they landed correctly.
Notice the tiers, though: two of the five biggest pots of the month were P-tier matches — a tier that accounted for 2.0% of all matches. Two hundred fifty-seven matches produced two of July’s five biggest pots. Which brings us to the genuinely interesting thing we found.
These are the longest streaks we observed on the board during July:
| Fighter | Win Streak | Tier | July Appearances |
|---|---|---|---|
| The world of arien | 103 | X | 1 |
| Ryousan gata patchouli | 78 | X | 1 |
| Honki-ultraman | 47 | X | 2 |
| Flea | 39 | X | 2 |
| Bloodbane | 32 | S | 3 |
| Fighter | Loss Streak |
|---|---|
| Guntank | -27 |
| Hakolen-sbs EX2 | -21 |
| Jackie chan | -21 |
| Simon belmont | -19 |
| Kiki de appendix | -18 |
An important honesty note on this table, because the numbers look wilder than they are: these are career streak counters, not July runs. The world of arien appeared on the SaltyBet card exactly once in all of July — and won, moving its all-time run to 103. It did not win 103 matches in July. It won one, on top of a streak it has been quietly building for months. Same story for Ryousan gata patchouli (78, one appearance — the same run we flagged back in May) and Flea (39, two appearances).
That pattern is the whole point, and it’s a real thing about SaltyBet rather than a footnote: every fighter in the top five is X-tier or S-tier, and they appear almost never. X-tier was 2.5% of July’s matches. These are the rare, dominant slot-fillers the matchmaker rolls out occasionally, and when it does, they win. Their streaks grow slowly precisely because they play so little. (Stat-nerd note: we lean on SaltyBet’s own per-match streak counter here rather than recounting from our archive, since our records only hold a slice of these rare fighters’ history — the counter is the authoritative number.)
On the other end, Guntank hit its 27th straight loss. Twenty-seven. Somewhere out there is a bettor still loyally riding Guntank, and we salute the commitment to the bit. :)
SaltyTrack’s model generated a prediction for every one of July’s 12,652 matches — 100% coverage, no gaps. Here’s how each confidence level performed:
| Confidence | Matches | Accuracy |
|---|---|---|
| Lock | 2 | 100% |
| High | 2,883 | 91.6% |
| Medium | 5,052 | 79.8% |
| Low | 4,715 | 59.8% |
| Overall | 12,652 | 75.0% |
The model hit 75.0% overall and the crowd hit 69.0% — a 6.1-point edge, or 768 additional correct calls across the month. For context on the trend:
| Month | Matches | SaltyTrack | Crowd | Edge |
|---|---|---|---|---|
| May 2026 | 12,873 | 75.1% | 68.4% | +6.7 |
| June 2026 | 12,192 | 73.4% | 68.5% | +4.9 |
| July 2026 | 12,652 | 75.0% | 69.0% | +6.1 |
July was a bounce-back after June’s softer month, essentially matching May’s model performance. Worth flagging honestly, though: the crowd also had its best month of the three at 69.0%. We didn’t gain ground because the crowd got worse — we gained it by getting better while the crowd did too.
The story, as always, is where the confidence lives. High-confidence calls hit 91.6% across 2,883 matches — that’s the band where the edge is real and the picks are worth backing. Medium (79.8%) was solid. Low (59.8%) is, by design, barely better than a coin toss — when our model says “Low,” it’s telling you it genuinely doesn’t know, and you should size your Salty Bucks accordingly. Note that Low was 37% of the whole month — SaltyBet throws a lot of genuinely unknowable matches at you, and a model that pretended otherwise would be lying. Lock fired just twice all month and hit both. For the full picture, here’s our deep dive on how SaltyTrack predicts fights.
The crowd bets biggest on P-tier — and it’s also the tier the crowd reads best. This is the finding that surprised us most this month, and it corrects something we’ve implied in past reports. P-tier drew July’s fattest average pots ($11.28M, nearly double A-tier). We’d assumed that was the crowd overcommitting to a tier it barely sees. It isn’t. Here’s crowd accuracy by tier:
| Tier | Crowd Accuracy | SaltyTrack |
|---|---|---|
| P | 79.0% | 85.2% |
| X | 78.5% | 86.4% |
| S | 70.2% | 76.4% |
| B | 69.6% | 73.9% |
| A | 66.7% | 73.4% |
The crowd is 12.3 points more accurate in P-tier than in A-tier. The reason, we think, is that P-tier and X-tier matches are frequently extreme mismatches — a joke character against something that actually functions — so the correct answer is obvious to everyone watching. The crowd bets big there because it’s confident, and it’s confident because it’s usually right. A-tier is the hard tier: deep, balanced, and full of genuinely close matchups. That’s where the crowd struggles, and it’s where our biggest edge lives too (+6.7 points).
We liked this one enough to give it its own piece: P-Tier: The Wild West of SaltyBet runs the same test across all of 2026 — 1,780 P-tier matches — and the ordering holds, so this isn’t a one-month artifact.
By our Chaos Score, July 12 was the wildest day of the month (score 231.5), edging out July 1 (228.3) and July 11 (220.3). What makes a day score high isn’t just upsets — it’s confident calls going wrong. July 12 saw 14 high-confidence predictions miss in a single day. The calmest day was July 15 (112.1), with only 3 high-confidence misses. Chaos isn’t about how many favorites lose; it’s about how many sure things do.
223 fighters climbed out of P-tier into B during July — while 104 went the other direction. Net, that’s 119 characters escaping the basement. We wrote about this exact revolving door back at the start of the month, and the headline finding still holds: a fresh promotion is not a buy signal.
How many SaltyBet matches happened in July 2026?
SaltyTrack recorded 12,652 matches between July 1 and July 31, 2026, with about $81.8 billion in total Salty Bucks wagered across 7,543 unique fighters — roughly 408 matches per day, at an average pot of $6.47 million.
Who was the top performer on SaltyBet in July 2026?
Master marco led all fighters at a perfect 10-0 in S-tier. Three fighters finished 9-0: Guts (S-tier), Nine the phantom (S-tier), and Lt.kurosawa EX (A-tier).
What was the biggest upset on SaltyBet in July 2026?
On July 31, Gomora ug beat Kimexemaru in an S-tier match at a 37.5x bet ratio — roughly 97% of the pot was on Kimexemaru. The costliest upset was July 5, when Kitana beat Bomber man at 24.1x inside a $54.1 million pot.
What was the biggest pot on SaltyBet in July 2026?
$372.1 million in a B-tier match between Vicviper and Green lantern gg on July 27, won by Green lantern gg. All five of July’s biggest pots were won by the betting favorite.
How accurate were SaltyTrack’s predictions in July 2026?
SaltyTrack’s model hit 75.0% overall across all 12,652 matches, beating the crowd’s 69.0% by 6.1 points. High-confidence predictions hit 91.6%, Medium hit 79.8%, and Low hit 59.8%.
Which SaltyBet tier does the crowd predict best?
P-tier, at 79.0% crowd accuracy in July 2026 — followed closely by X-tier at 78.5%. The crowd was least accurate in A-tier at 66.7%, SaltyBet’s deepest and most competitive tier.
Want SaltyTrack’s confidence calls on every match before you bet? The SaltyTrack Chrome Extension overlays Lock, High, Medium, and Low predictions right on the SaltyBet Twitch channel — plus fighter stats, head-to-head records, and streak alerts. It’s free, no account, no nonsense — and in July, those calls beat the crowd by 6.1 points.
SaltyBet uses virtual currency only. No real money is wagered or exchanged. SaltyTrack is not affiliated with SaltyBet.