The September archive shows 12,505 recorded matches and $61.15 billion wagered. The crowd-favorite side won 70.0% of the time, while SaltyTrack models picked the winner 67.4% of the time. Beautiful family, welcome back. The chat gets the bragging rights this time, and Vanguard zero walked through an S-tier door with just 1.9% of the money to turn a $58.2 million pot into a 51.8-to-1 upset.[1]
One honest footnote before we roll the tape: September 29 has only 77 recorded outcomes, all before 04:01 UTC. No later results appear in the available archive. This is our best read of the records available today, not a complete September ledger. We’ll refresh the totals if more source data turns up.
| Metric | Recorded value |
|---|---|
| Matches in local archive | 12,505 |
| Recorded Salty Bucks wagered | $61.15 billion |
| Average pot | $4.89 million |
| Distinct fighter names | 7,696 |
| Red wins | 6,122 (49.0%) |
| Blue wins | 6,383 (51.0%) |
| Match format | 10,844 matchmaking / 1,661 tournament |
| Collection window | September 1–30, 2026 (UTC) |
7,696 fighter names across 12,505 matches. That’s 25,010 fighter slots, or just over three appearances per name on average. SaltyBet’s roster is still an ocean; most characters drift through the month a few times and then the matchmaker spins the wheel again. New to the churn? Our Beginner’s Guide lays out how the matches and tiers work.[1]
Blue finished ahead 51.0% to 49.0%, a difference of 261 matches. We say this every month: the red/blue split is a coin flip wearing a different jersey. September was no exception.
| Tier | Recorded matches | Share | Average pot |
|---|---|---|---|
| A | 5,289 | 42.3% | $4.36M |
| S | 3,981 | 31.8% | $5.11M |
| B | 2,618 | 20.9% | $5.37M |
| X | 338 | 2.7% | $5.79M |
| P | 279 | 2.2% | $6.24M |
A-tier and S-tier together made up 74.1% of the recorded card; that’s where most of the action lives. P-tier was the oddball on the right side of the table: just 279 matches, but the highest average pot at $6.24 million. That’s a fun September result, and the sample is a lot smaller than A or S, so it is not a lifetime ranking.
We used a ten-match floor so one lucky afternoon doesn’t crown a champion. Suwako ai went a perfect 10–0, the only fighter in this top group to clear the bar without a loss. Shindo tamaki led this leaderboard in appearances, going 12–1 across 13 matches.[1]
| Fighter | Record | Win rate |
|---|---|---|
| Suwako ai | 10–0 | 100.0% |
| Shindo tamaki | 12–1 | 92.3% |
| Johnny(ac) | 11–1 | 91.7% |
| Bloodbane | 10–1 | 90.9% |
| Aaaaaaaaaaaahhhhhh!!! | 9–1 | 90.0% |
| Guts | 9–1 | 90.0% |
| Justice | 9–1 | 90.0% |
| Old mechagodzilla | 9–1 | 90.0% |
Four-way tie at 90%: Aaaaaaaaaaaahhhhhh!!!, Guts, Justice, and Old mechagodzilla each finished 9–1. That’s a lot of excellent September work for a leaderboard that still has one clear headline: Suwako ai did not lose.
And on the other end of the standings, Pogo went 0–8. A rough month is a rough month; it isn’t a lifetime scouting report. With this many fighters and a small monthly sample, the matchmaker can hand you a brutal run just as quickly as it hands someone else a perfect one.
These are the matches where the crowd piled money on the favorite, but the favorite lost. We rank by how many Salty Bucks sat on the losing side for every dollar on the winner.
Vanguard zero beat Neff in S-tier on September 5 with just 1.9% of the $58.2 million pot behind it. The crowd had stacked about $57.1 million on Neff and just over $1.1 million on Vanguard zero. That’s a 51.8-to-1 favorite-to-winner split, and a proper September-sized gasp.[1]
| Date | Underdog (Winner) | Favorite (Loser) | Tier | Favorite stake per $1 on winner | Pot |
|---|---|---|---|---|---|
| Sep 5 | Vanguard zero | Neff | S | 51.8-to-1 | $58.2M |
| Sep 20 | Maccho | Goenitz EX3 | S | 22.9-to-1 | $1.4M |
| Sep 26 | Toudou | Shimo/grand order | X | 20.5-to-1 | $1.1M |
| Sep 27 | Relius clover | Opirus element | S | 18.5-to-1 | $0.6M |
| Sep 20 | Custom ssfiv oni | Jack the ripper | B | 17.4-to-1 | $0.5M |
The 51.8-to-1 result had both things an upset story wants: a truly lopsided crowd and a real pot behind it. The next four on the list came in smaller pools. The chaos was still great, with a much smaller audience holding its breath.
The largest recorded pot was $657.7 million on September 21, when Patoelietoe defeated Vyres in a B-tier match.[1] The next-largest was $272.7 million, less than half as large.
| Date | Matchup | Winner | Tier | Pot |
|---|---|---|---|---|
| Sep 21 | Vyres vs Patoelietoe | Patoelietoe | B | $657.7M |
| Sep 17 | Kurorim vs Goku sb1 | Kurorim | B | $272.7M |
| Sep 7 | Ikki vs Kiki de appendix | Ikki | B | $191.7M |
| Sep 18 | Vava vs Albert roseman | Albert roseman | B | $69.5M |
| Sep 21 | Mettaton neo vs Naked snake | Naked snake | P | $59.4M |
B-tier took three of the five biggest pots. The top one was in a different universe from the rest: $657.7 million on Patoelietoe vs. Vyres, then a steep drop to $272.7 million. B-tier doesn’t always get the headlines, but the whales showed up there in September. For a tour of the tier system, start here.
Across the 12,505 recorded matches, SaltyTrack models picked the winner 8,423 times (67.4%), while the side with more money at the end won 8,754 times (70.0%). The crowd gets the September bragging rights by 2.6 points.[1] That crowd number is a hindsight comparison based on final betting totals; it isn’t a signal we can know at the start of every match.
| Prediction confidence | Correct | Matches | Accuracy |
|---|---|---|---|
| High | 2,350 | 2,857 | 82.3% |
| Medium | 3,326 | 4,925 | 67.5% |
| Low | 2,747 | 4,723 | 58.2% |
| Overall | 8,423 | 12,505 | 67.4% |
| Crowd favorite | 8,754 | 12,505 | 70.0% |
High-confidence predictions were SaltyTrack models’ strongest band at 82.3%; low-confidence predictions landed at 58.2%. That separation is exactly why confidence matters: a “Low” call is a signal of uncertainty, not a promise that SaltyTrack models have a strong side. For how those confidence labels work, see How SaltyTrack Predicts Fights.
And the tier split is where the report gets weird in a useful way:
| Tier | SaltyTrack | Crowd favorite | Matches |
|---|---|---|---|
| A | 65.9% | 67.1% | 5,289 |
| S | 70.0% | 71.7% | 3,981 |
| B | 65.6% | 72.5% | 2,618 |
| X | 73.7% | 70.4% | 338 |
| P | 66.7% | 76.3% | 279 |
SaltyTrack models’ only tier-level edge was X. The crowd was strongest in P at 76.3%, though that’s only 279 matches, and B had the widest gap: 72.5% for the crowd against 65.6% for SaltyTrack models. P-tier’s crowd strength echoes what we found in our P-tier deep dive; B-tier is September’s new wrinkle.[1] We’ll rerun the whole table if September 29 coverage changes.
September 2 was SaltyTrack models’ roughest high-confidence night, with 26 misses in one day.[1] Even the strong calls have a bad night now and then; the useful question is whether the confidence bands still separate strong reads from coin flips over thousands of matches. For September, they did: High landed at 82.3%, while Low was 58.2%. Our prediction explainer walks through what those labels mean.
How many SaltyBet matches were recorded in September 2026?
The SaltyTrack archive contains 12,505 matches dated September 1–30, 2026. September 29 is incomplete in the available production and local records: 77 canonical matches appear before 04:01 UTC, with no later outcomes recorded.
Who was September’s top undefeated fighter?
Suwako ai went 10–0 in the local September sample, the best undefeated record among fighters with at least ten recorded matches.
What was the biggest SaltyBet upset in September?
On September 5, Vanguard zero defeated Neff in S-tier despite attracting roughly 1.9% of the $58.2 million pot, a 51.8-to-1 mismatch in betting support.
How accurate was SaltyTrack in September 2026?
SaltyTrack models picked the winner in 8,423 of 12,505 recorded matches, or 67.4%. High-confidence predictions were 82.3% accurate, medium 67.5%, and low 58.2%.
Did SaltyTrack or the crowd predict more winners?
The side with the larger final bet total won 70.0% of recorded September matches, compared with 67.4% for SaltyTrack models. September 29 has 77 recorded outcomes, all before 04:01 UTC, with no later results available.
Want SaltyTrack’s confidence calls on every match before you bet? The SaltyTrack Chrome Extension overlays Lock, High, Medium, and Low predictions on the SaltyBet Twitch channel, alongside fighter stats, head-to-head records, and streak alerts. It’s free, no account, no nonsense.
SaltyBet uses virtual currency only. No real money is wagered or exchanged. SaltyTrack is not affiliated with SaltyBet.