Here’s a situation for the number crunchers and risk junkies:
We are the Sheriff (6 man arena mode) with a weak character, zero defensive cards — only a Volcanic and a fistful of Bang!s.
On our first turn we can shoot someone down right away.
We estimate a 60% chance to hit an Outlaw.
If we do, the game shifts to a 5-player setup — historically very Sheriff-favored, but let’s stay conservative and call it 50% winrate if we hit.
If we do nothing and just sit with no defense, we rate our winrate at ~20%.
If we shoot and accidentally kill a Deputy\Rene (40% chance) — we auto-lose: 0%.
So:
Do nothing: ~20% winrate
Shoot: ??
Net EV Formula
We’re not just interested in raw outcome — we want the net value compared to doing nothing:
Net EV=(0.6×0.5)+(0.4×0)−0.2=0.3−0.2=0.1
So pulling the trigger adds a +10% net gain compared to sitting tight and hoping for mercy. In realistic scenarion even better (realistic WR for team sheriff in 5 man state about 60%), and NET EV will be +16%
Takeaway
It’s ugly. It’s risky. It might kill your own team.
But mathematically, if you’re Sheriff with no defense and a bad hero — taking the shot is +EV.
Sometimes your only armor is raw aggression.
Scenario:
Imagine you’re the Deputy, sitting directly after the Sheriff. Your main job is to act as the classic “meat shield” — basically, your HP is your Sheriff’s spare life bar.
The table is 6 players. The Sheriff isn’t exactly a powerhouse — not Bart, not Janet, maybe something mid-tier or weak.
If you do nothing, the setup is decent enough to expect an average Sheriff winrate around 40% — that’s about what we see on BoardGameArena for good Sheriffs in a 6-player lineup. (If sheriff bad hero, or bad player + bad hero, this WR can be worse)
Now, you’ve got a shot: you can try to kill the next player after you (we have this opportunity to kill) There’s a 75% chance you hit an Outlaw. If you do, the table drops to 5 players — historically, that bumps the Sheriff’s winrate up hard. AVTactics stats say the Sheriff’s team often hits 60% in this 5-player state.
There’s also a 25% chance you accidentally kill the Renegade. That’s bad, but not quite instant loss — let’s assume the Sheriff’s odds drop to 5% if that happens.
So the question: Is this shot worth it?
Net EV Calculation
Here’s the clean math:
Net EV=(0.75×0.60)+(0.25×0.05)−0.40=0.45+0.0125-0.40=0.0625
Takeaway
Your shot adds +6.25% net value compared to just staying passive and hoping the meat shield plan holds.
Yes, there’s risk — a dead Renegade can backfire — but real data says the Sheriff’s odds skyrocket when the table shrinks to 5.
BIG UPDATE
I collected a database in which I researched over 9,000 samples of Sheriff games. I’ll make a separate post about the database methodology.
In this post specifically, I want to update some numbers that were previously based mostly on intuition. I used to assume that the average Sheriff win rate for an average player was around 40%. But reality turned out to be very different.
I was using my own Sheriff win rate — around 40% — as a reference point. But for the average Bang! player, the Sheriff win rate is actually around 31%. (from this database)
So the situation discussed in that post, in a vacuum and without accounting for characters, looks even better than expected. If we take the real average win rate of 31%, and after the kill that win rate becomes close to 50%, then random killing becomes even more profitable than I originally thought. (for the random kill from helper)
In the actual data, the average Sheriff win rate rises to about 45% after the first Outlaw dies. But here I’m using 50%, because the Sheriff’s win rate after the first dead Outlaw is calculated across all situations where an Outlaw died. In those games, the Sheriff may have been damaged to different degrees.
But the situation where the Sheriff remains at 4-5 HP -- the win rate is quite high.
So the jump from the average Sheriff win rate of 31% to around 50% makes the situation with a random kill even more profitable than expected. (for the random kill by vice)
I’m also attaching research on characters and their win rates. When a character has more than 500 games, that win rate is already quite close to reality. (FOR THE AVERAGE BANG PLAYER) With 500+ games per character, the margin of error is roughly within a 3% range.
Now about sheriff"s random kill, for example, if you end up playing Sheriff as Willy, whose win rate appears to be around 20% — maybe 25% at most, in my estimation — then your goal should be very clear. When you pick Willy, you should always have one main task: if you draw a bunch of Bang! cards on your first turn, you should try to kill someone.
That is your game plan. You pick Willy with the strict plan of trying to kill somebody.
I also want to point out that my assumption that the win rate drops to around 20% when you have no defensive cards turned out to be close to reality.
At the moment, the actual number is around 22% if you have zero defensive cards.
And if you are playing a character whose Sheriff win rate is already only around 20–25%, then that drop can be even more devastating.
We are the Sheriff (6 man arena mode) with a weak character, zero defensive cards — only a Volcanic and a fistful of Bang!s.
On our first turn we can shoot someone down right away.
We estimate a 60% chance to hit an Outlaw.
If we do, the game shifts to a 5-player setup — historically very Sheriff-favored, but let’s stay conservative and call it 50% winrate if we hit.
If we do nothing and just sit with no defense, we rate our winrate at ~20%.
If we shoot and accidentally kill a Deputy\Rene (40% chance) — we auto-lose: 0%.
So:
Do nothing: ~20% winrate
Shoot: ??
Net EV FormulaWe’re not just interested in raw outcome — we want the net value compared to doing nothing:
Net EV=(0.6×0.5)+(0.4×0)−0.2=0.3−0.2=0.1
So pulling the trigger adds a +10% net gain compared to sitting tight and hoping for mercy. In realistic scenarion even better (realistic WR for team sheriff in 5 man state about 60%), and NET EV will be +16%
TakeawayIt’s ugly. It’s risky. It might kill your own team.
But mathematically, if you’re Sheriff with no defense and a bad hero — taking the shot is +EV.
Sometimes your only armor is raw aggression.
Scenario:Imagine you’re the Deputy, sitting directly after the Sheriff. Your main job is to act as the classic “meat shield” — basically, your HP is your Sheriff’s spare life bar.
The table is 6 players. The Sheriff isn’t exactly a powerhouse — not Bart, not Janet, maybe something mid-tier or weak.
If you do nothing, the setup is decent enough to expect an average Sheriff winrate around 40% — that’s about what we see on BoardGameArena for good Sheriffs in a 6-player lineup. (If sheriff bad hero, or bad player + bad hero, this WR can be worse)
Now, you’ve got a shot: you can try to kill the next player after you (we have this opportunity to kill) There’s a 75% chance you hit an Outlaw. If you do, the table drops to 5 players — historically, that bumps the Sheriff’s winrate up hard. AVTactics stats say the Sheriff’s team often hits 60% in this 5-player state.
There’s also a 25% chance you accidentally kill the Renegade. That’s bad, but not quite instant loss — let’s assume the Sheriff’s odds drop to 5% if that happens.
So the question: Is this shot worth it?
Net EV CalculationHere’s the clean math:
Net EV=(0.75×0.60)+(0.25×0.05)−0.40=0.45+0.0125-0.40=0.0625
TakeawayYour shot adds +6.25% net value compared to just staying passive and hoping the meat shield plan holds.
Yes, there’s risk — a dead Renegade can backfire — but real data says the Sheriff’s odds skyrocket when the table shrinks to 5.
BIG UPDATE
I collected a database in which I researched over 9,000 samples of Sheriff games. I’ll make a separate post about the database methodology.
In this post specifically, I want to update some numbers that were previously based mostly on intuition. I used to assume that the average Sheriff win rate for an average player was around 40%. But reality turned out to be very different.
I was using my own Sheriff win rate — around 40% — as a reference point. But for the average Bang! player, the Sheriff win rate is actually around 31%. (from this database)
So the situation discussed in that post, in a vacuum and without accounting for characters, looks even better than expected. If we take the real average win rate of 31%, and after the kill that win rate becomes close to 50%, then random killing becomes even more profitable than I originally thought. (for the random kill from helper)
In the actual data, the average Sheriff win rate rises to about 45% after the first Outlaw dies. But here I’m using 50%, because the Sheriff’s win rate after the first dead Outlaw is calculated across all situations where an Outlaw died. In those games, the Sheriff may have been damaged to different degrees.
But the situation where the Sheriff remains at 4-5 HP -- the win rate is quite high.
So the jump from the average Sheriff win rate of 31% to around 50% makes the situation with a random kill even more profitable than expected. (for the random kill by vice)
I’m also attaching research on characters and their win rates. When a character has more than 500 games, that win rate is already quite close to reality. (FOR THE AVERAGE BANG PLAYER) With 500+ games per character, the margin of error is roughly within a 3% range.
Now about sheriff"s random kill, for example, if you end up playing Sheriff as Willy, whose win rate appears to be around 20% — maybe 25% at most, in my estimation — then your goal should be very clear. When you pick Willy, you should always have one main task: if you draw a bunch of Bang! cards on your first turn, you should try to kill someone.
That is your game plan. You pick Willy with the strict plan of trying to kill somebody.
I also want to point out that my assumption that the win rate drops to around 20% when you have no defensive cards turned out to be close to reality.
At the moment, the actual number is around 22% if you have zero defensive cards.
And if you are playing a character whose Sheriff win rate is already only around 20–25%, then that drop can be even more devastating.
