Bo1 vs Bo3 Upsets: How Series Length Protects Favourites (Model)
A model study of how Bo1, Bo3 and Bo5 formats turn the same per-map edge into very different upset rates and fair underdog prices.
Model studyEvery number on this page is calculated from the rules of the sport and stated assumptions. These are not historical match results.
Every esports bettor has felt it: the clear favourite drops a single map and the Bo1 is gone. I wanted to put numbers on that feeling, so this is a model study. It does not use match records. It is an exact calculation of how series length changes the result when one team has a fixed chance of winning each map.
The headline answer is that longer series protect favourites, but only once the edge is real. In this model a 60% map favourite is upset 40.0% of the time in a Bo1, 35.2% in a Bo3 and 31.7% in a Bo5. At an 80% map edge the Bo3 upset rate falls to 10.4%. I use these numbers as a baseline before I look at form and map pools for my free esports picks page.
Period 1Upset rates fall faster in longer series as edges grow
This chart plots the underdog's chance of winning the series against the favourite's map win chance, with one line for each format. At 50% all three lines start together, because a coin flip stays a coin flip however many times you toss it. As the favourite's edge grows, the lines fan out, and the Bo5 line drops away fastest.
At 70% per map the model upset rate is 30.0% in a Bo1, 21.6% in a Bo3 and 16.3% in a Bo5. At 85% the Bo5 upset rate is just 2.7%, against 15.0% in a Bo1. The format does as much work as the edge itself.
View the data: underdog series win: bo1 vs bo3 vs bo5
| Bo1 | Bo3 | Bo5 | |
|---|---|---|---|
| 50% | 50.0% | 50.0% | 50.0% |
| 55% | 45.0% | 42.5% | 40.7% |
| 60% | 40.0% | 35.2% | 31.7% |
| 65% | 35.0% | 28.2% | 23.5% |
| 70% | 30.0% | 21.6% | 16.3% |
| 75% | 25.0% | 15.6% | 10.4% |
| 80% | 20.0% | 10.4% | 5.8% |
| 85% | 15.0% | 6.1% | 2.7% |
The part I find most useful is the bottom-left corner. At 55% the model gives 45.0%, 42.5% and 40.7%, so a slight favourite gets almost no protection from a longer series. A marginal edge in a Bo5 is still close to a coin flip, and the market is not always priced that way.
Period 2Bo3 turns a modest map edge into a stronger series favourite
Here the same map win chance sits at both ends of each dumbbell: the Bo1 series win on one side and the Bo3 series win on the other. In a Bo1 the two numbers are identical by definition. The Bo3 stretches them apart.
In this model a 60% map favourite becomes a 64.8% series favourite in a Bo3, and a 70% map favourite becomes 78.4%. A 75% map favourite wins 84.4% of model Bo3 series against 75.0% of Bo1 matches. The gap is widest in the middle of the range, where most real pricing decisions sit.
View the data: favourite series win: bo1 vs bo3
| Bo1 | Bo3 | |
|---|---|---|
| 55% map win (model) | 55.0% | 57.5% |
| 60% map win (model) | 60.0% | 64.8% |
| 65% map win (model) | 65.0% | 71.8% |
| 70% map win (model) | 70.0% | 78.4% |
| 75% map win (model) | 75.0% | 84.4% |
| 80% map win (model) | 80.0% | 89.6% |
When I see two teams priced almost the same across formats, this chart is my first check. If the per-map edge has not changed, the Bo3 price should be noticeably shorter for the favourite than the Bo1.
Period 3Fair underdog odds lengthen sharply from Bo1 to Bo5
The table converts the model upset chances into fair decimal odds, which are simply one divided by the probability, with no bookmaker margin. At a 60% map edge the underdog's fair price is 2.50 in a Bo1, 2.84 in a Bo3 and 3.15 in a Bo5.
Further down the table the spread becomes dramatic. At 70% per map the underdog's fair odds lengthen to 4.63 in a Bo3 and 6.13 in a Bo5, and at 80% the Bo5 underdog is a fair 17.27. These are the numbers a price has to beat before an underdog bet has any value at all.
| Map win | Bo1 upset | Bo3 upset | Bo5 upset | Bo3 dog fair | Bo5 dog fair |
|---|---|---|---|---|---|
| 55% | 45.0% | 42.5% | 40.7% | 2.35 | 2.46 |
| 60% | 40.0% | 35.2% | 31.7% | 2.84 | 3.15 |
| 65% | 35.0% | 28.2% | 23.5% | 3.55 | 4.25 |
| 70% | 30.0% | 21.6% | 16.3% | 4.63 | 6.13 |
| 75% | 25.0% | 15.6% | 10.4% | 6.40 | 9.66 |
| 80% | 20.0% | 10.4% | 5.8% | 9.62 | 17.27 |
The main caveat applies most here. Real maps are not independent coin flips with a fixed chance. Vetoes, map pools, momentum, form and roster changes all move the true edge, and Bo1 events often use a different veto that can change the edge itself. Treat these prices as what the format does to a given strength gap, not as a forecast for any match.
Period 4Bo3 and Bo5 need a far smaller map edge
This chart works backwards. It asks how strong a team must be on each map to reach a target series win of 70%, 80% or 90%. In a Bo1 the answer equals the target, because the map is the series.
Longer formats lower the bar. To be an 80% series favourite in this model, a team needs 80.0% per map in a Bo1 but only 71.3% in a Bo3 and 67.3% in a Bo5. For a 90% series favourite the model needs 80.4% per map in a Bo3 and 75.3% in a Bo5, and a 70% series favourite needs 63.7% and 61.0%.
View the data: map win chance needed for a series win target
| Model: map win chance that gives a 70%, 80% or 90% series win | |
|---|---|
| Bo1 ยท 70% series | 70.0% |
| Bo3 ยท 70% series | 63.7% |
| Bo5 ยท 70% series | 61.0% |
| Bo1 ยท 80% series | 80.0% |
| Bo3 ยท 80% series | 71.3% |
| Bo5 ยท 80% series | 67.3% |
| Bo1 ยท 90% series | 90.0% |
| Bo3 ยท 90% series | 80.4% |
| Bo5 ยท 90% series | 75.3% |
I use this when a series price looks short. If the market implies around an 80% Bo3 favourite, I ask whether I believe that team wins roughly 71.3% of its maps against this opponent on this pool. If my map read says otherwise, the series price is wrong too.
Full-time verdict
Estimate the per-map edge, then let the format set the series price. In this model a 70% map edge becomes 78.4% in a Bo3, not 70%.
At a 55% map edge the model Bo5 upset rate is still 40.7%, so a longer series does little to protect a marginal favourite. Underdogs in these spots deserve a serious look.
Before backing a Bo3 underdog against a 60% map favourite, compare the price with the model's fair 2.84. Anything shorter needs a specific reason, such as a veto or roster edge the model ignores.
Questions bettors ask
Are Bo1 matches more likely to produce upsets than Bo3?
Yes, in this model. A 60% map favourite loses a Bo1 40.0% of the time but a Bo3 only 35.2%, and the gap widens as the edge grows.
How often does the underdog win a Bo5 in esports?
In this model it depends heavily on the edge. Against a 70% map favourite the Bo5 upset rate is 16.3%, and against an 85% map favourite it is just 2.7%.
What are fair odds for a Bo3 underdog?
The model's fair underdog price is 2.84 against a 60% map favourite and 4.63 against a 70% map favourite. These carry no bookmaker margin and assume each map is independent.
I'm Charlotte Reed, based in London, and I write the eSports game predictions at correct-score.vip โ covering the major titles where form and meta shift fast.
View profile โ
