Which Day of an Esports Event Has the Most Upsets? A Model
A simulated six-day event shows upsets peak on the last group day, not the opening one, and that longer series quietly shield the rated favourite.
Written by Camille LefรจvreeSports Betting Tips specialist
- Sport
- Esports
- Question
- In a model six-day esports event with Bo1 group days and Bo3/Bo5 playoff days, on which day does the rated favourite lose most often?
- Sample
- Model study: 40,000 simulated 16-team events (Bo1 groups, Bo3/Bo5 playoffs); no historical match records used
- Upset rate on Day 1, Bo1 group games (model)31.9%
- Upset rate in the Bo5 final (model)30.9%
- All group-day upsets (model)33.8%
- All playoff-day upsets (model)32.9%
Model studyEvery number on this page is calculated from the rules of the sport and stated assumptions. These are not historical match results.
Does the day of an esports event change how often the favourite falls? I built this as a model study, not a trawl of past results: 40,000 simulated 16-team events with Bo1 group days followed by Bo3 and Bo5 playoffs. The model tracks how often the rated favourite loses on each of the six days.
The headline in this model is that Day 3 is the upset day, at 38.7%, and the Bo5 final is the calmest at 30.9%. This sits alongside my esports betting tips today, but it answers a calendar question rather than a matchup one. Treat every figure below as the model's output under stated assumptions, never as a historical rate.
Window 1The last group day breaks favourites most often
Across all simulated matches the rated favourite loses 33.6% of the time in this model. The first two group days sit below that, at 31.9% and 30.8%. Then Day 3 jumps to 38.7%, the highest of the whole event.
Once the playoffs start, the model settles: 33.3% in the Bo3 quarter-finals, 33.2% in the semi-finals and 30.9% in the Bo5 final. The group stage as a whole runs at 33.8% against 32.9% for the playoff days, a gap of 0.9 points.
View the data: upset rate by day of a six-day event
| Model: rated favourite loses; 16 teams, Bo1 groups, Bo3/Bo5 playoffs | |
|---|---|
| Day 1 ยท group Bo1 | 31.9% |
| Day 2 ยท group Bo1 | 30.8% |
| Day 3 ยท group Bo1 | 38.7% |
| Day 4 ยท quarter-finals Bo3 | 33.3% |
| Day 5 ยท semi-finals Bo3 | 33.2% |
| Day 6 ยท final Bo5 | 30.9% |
| All matches | 33.6% |
What I take from this is that the chaos most bettors expect from opening-day Bo1s doesn't appear on Day 1 in the model. It shows up later in the group stage, and the reason is the schedule rather than the format.
Window 2Day 3 pairs the closest teams in the field
The table explains the Day 3 spike. On Day 1 the rated favourite's true single-map chance averages 67.9%, and on Day 2 it is 69.2%. By Day 3 it drops to 61.3%, because the model's schedule saves the meeting of each group's top two seeds for last.
That turns into prices. The model's fair upset odds are 3.13 on Day 1 but only 2.58 on Day 3. For the quarter-finals they are 3.00, and for the final 3.24. These fair odds carry no bookmaker margin, so they are reference points rather than prices you will actually be offered.
| Day | Matches | Fav map chance | Upset rate | 95% range (pts) | Fair upset odds |
|---|---|---|---|---|---|
| Day 1 | 8 | 67.9% | 31.9% | ยฑ0.2 | 3.13 |
| Day 2 | 8 | 69.2% | 30.8% | ยฑ0.2 | 3.25 |
| Day 3 | 8 | 61.3% | 38.7% | ยฑ0.2 | 2.58 |
| Day 4 QF | 4 | 62.0% | 33.3% | ยฑ0.2 | 3.00 |
| Day 5 SF | 2 | 61.9% | 33.2% | ยฑ0.3 | 3.02 |
| Day 6 final | 1 | 61.6% | 30.9% | ยฑ0.5 | 3.24 |
The playoff days have mismatches about as tight as Day 3, with favourite map chances of 62.0%, 61.9% and 61.6%, yet they produce fewer upsets. The next chart shows why.
Window 3Longer series absorb most of the playoff upset risk
Here the model replays the same matchups as if each were a single map. On the group days nothing changes, because they already are Bo1: Day 1 is 32.1% by construction against 31.9% in the simulation, and that difference is simulation noise.
The playoffs are where the format matters. The quarter-finals would run at 38.0% as Bo1s instead of 33.3% as Bo3s. The final's matchups would produce an upset 38.4% of the time as a Bo1, and the Bo5 cuts that to 30.9%.
View the data: same matchups as a bo1 vs the real format
| If played Bo1 | Actual format | |
|---|---|---|
| Day 1 (320,000 sim. matches) | 32.1% | 31.9% |
| Day 2 (320,000 sim. matches) | 30.8% | 30.8% |
| Day 3 (320,000 sim. matches) | 38.7% | 38.7% |
| Day 4 QF (160,000 sim. matches) | 38.0% | 33.3% |
| Day 5 SF (80,000 sim. matches) | 38.1% | 33.2% |
| Day 6 final (40,000 sim. matches) | 38.4% | 30.9% |
So in this model playoff favourites aren't facing easier opponents. Day 3 and the final have near-identical map edges, and the extra maps are what protect the stronger side.
Window 4Field depth moves every day, but the shape holds
The base model uses a rating spread of 0.6. With a tighter field at 0.4, Day 1 upsets rise to 38.1% and the final to 36.8%. With a wider field at 0.8, both Day 1 and the final fall to 27.5%.
In all three versions Day 3 stays the peak, reaching 43.1% in the tight field and 35.2% in the wide one. The level depends heavily on how even the event is, but the pattern by day stays the same.
View the data: upset rate by event day and field spread
| Spread 0.4 | Spread 0.6 (base) | Spread 0.8 | |
|---|---|---|---|
| Day 1 | 38.1% | 31.9% | 27.5% |
| Day 2 | 37.6% | 30.8% | 25.0% |
| Day 3 | 43.1% | 38.7% | 35.2% |
| Day 4 QF | 38.5% | 33.3% | 29.3% |
| Day 5 SF | 38.4% | 33.2% | 29.6% |
| Day 6 final | 36.8% | 30.9% | 27.5% |
My main caveat is that this is a model with assumed strengths and the same map edge on every day. Real Bo1 groups use different veto rules and map pools from playoff series, and teams may hide strategies early, which the model doesn't capture.
Best window
- Target the final group day
In this model Day 3 produces upsets 38.7% of the time with fair upset odds of 2.58, because the strongest pairs in each group meet then. That is the day to look hardest at underdog prices.
- Respect the Bo5 final
The same matchups would produce an upset 38.4% of the time as a Bo1, but the Bo5 brings the model rate down to 30.9%. Don't price a final upset as if it were a single map.
- Read the field first
A tight field at spread 0.4 pushes Day 1 upsets to 38.1%, while a wide one at 0.8 drops them to 27.5%. Judge how even the event is before trusting any day-based angle.
Questions bettors ask
Which day of an esports event has the most upsets?
In this model it is Day 3, the last Bo1 group day, where the rated favourite loses 38.7% of the time. The schedule pairs each group's top two seeds on that day, so the matchups are closest.
Are Bo1 group games more upset-prone than playoff series?
In the model, group days run at 33.8% upsets against 32.9% across the playoff days. The longer formats do most of the work: the quarter-finals would be 38.0% as Bo1s but are 33.3% as Bo3s.
How often does the favourite lose an esports final?
In the base model the Bo5 final produces an upset 30.9% of the time, with fair upset odds of 3.24. With a tighter field that rises to 36.8%. These are model figures, not historical results.
I'm Camille Lefรจvre, based in Lyon, and I write the esports betting tips at fixedmatches.vip โ covering the major titles where form and meta turn over fast.
View profile โ
