---
title: Bunker mode in Oracle
tags: Oracle, withdrawals
status: Draft
author: Raman Siamionau
created: 2022-10-21
updated: 2023-01-14
---

# Bunker mode

## Why bunker mode exists?

When slashing begins we are not sure how much ETH our validators will lost. All begins with small amounts and if there are a lot of slashed validators penalty may increase to full validator's balance.

## Economic


The most safest case is to enable bunker mode when slashing begins.
The most optimistic case is to enable bunker mode only when delta is negative. (That means are not expecting another loses and no additional slashing expected).

## Soft and Hard strategy

#### Strategy 1

def is_bunker_mode(
	lido_slashings_last_36d, 
	non_lido_slashings_last_36d, 
	median_rewards_per_day
):
	# calculate slashing losses by day based on current data 
	losses_for_next_days = get_penalties_by_day(
            lido_slashings_last_36d,
            non_lido_slashings_last_36d
        )
	
	# the minimum slashing duration is 36 days
	assert len(losses_for_next_days) >= 36
  
	for loss_per_day in losses_for_next_days: 
		if loss_per_day > median_rewards_per_day:
			return True
	
	return False

#### Strategy 2

def is_bunker_mode(
    lido_slashed_valudators_balance_sum,
    lido_income_for_a_day,
):
    if lido_income_for_a_day < 0:
        return True
    
    if lido_slashed_valudators_balance_sum > lido_income_for_a_day:
        return True
    
    return False

#### Strategy 3

def is_bunker_mode(...):
    if lido_income_for_a_day < 0:
        return True

    def risks(lido_validars_amount_slashed, total_validators_slashed_percent):
        return (e * max(total_validators_slashed_percent / 0.3, 0.01) + 1)**lido_validars_amount_slashed
    
    return risks(...) > 10

If 10% of all validators were slashed

If 30% of all validators were slashed