Let’s just round this up to 600k gas units. With ETH at $1802 this is what it cost with different gas prices
I am comfortable with covering gas up to 100 gwei which is roughly $108 and with BNT at .42 this is roughly 257 BNT or so. I suggest that the adjustment is made to be 300 BNT cap for now (a 3x increase).
The contract caller always pays for gas - Bancor never covers gas costs.
Bancor always receives a minimum 50% cut, and receives everything in excess of the cap.
When the gas price is high, there is currently a point at which the Fast Lane becomes unusable. For example right now, if gas costs more than 100 BNT in ETH to execute the arbitrage, the caller can’t execute the arb because the gas costs more than the maximum reward. During the recent activity spike, we saw gas costs of >500 BNT.
Here are a few examples for anyone reading this - because I think it may not be clear for a lot of people:
Bot finds an arb opportunity that would result in a profit of 300 BNT, and costs 80 BNT in gas. Bot executes the contract, pays 80 BNT worth of gas, and receives 100 BNT - the current max reward. Bancor burns its cut of 200 BNT.
Bot finds an arb opportunity that would result in a profit of 300 BNT, and costs 110 BNT in gas. Bot cannot execute, because the maximum reward is less than the gas cost of the transaction.
Proposed @ 2000 cap:
Bot finds an arb opportunity that would result in 600 BNT profit, and costs 250 BNT in gas. Bot executes and receives 300 BNT for a profit of 50 BNT. Bancor burns its cut of 300 BNT.
Bot finds an arb opportunity of 5000 BNT profit, and costs 2001 BNT in gas. Bot cannot execute because the maximum reward is less than the gas cost of the transaction.
The reason for this proposed change is simply to support periods of higher network activity. The caller still pays 100% of the gas cost and carries the risk of paying for failed transactions.
100 to 2000 is 1 order of magnitude
20 orders of magnitude would be 10,000,000,000,000,000,000,000 BNT. I bring this up because I would support either of these. Arb opportunities in the range of $1,000 are infrequent; opportunities in the range of $2,000 are less frequent still. Arb opportunities in the range of hextillions of dollars are fantasy. At some point there is no difference between the limit set, and not having a limit.
I have no strong feelings about the existence of the cap. In my view the forced 50% share is sufficient until evidence to the contrary is presented. The proposed 2,000 BNT limit is high enough that it no longer matters - which makes me happy.
Nice and necessary proposal.
Gas have been too high lately, and it is predictable that they will remain high during times of volatility.
Most of the time BNT rewards cannot cover gas costs. The contract has not been effectively invoked for a long time.
Raising rewards to 1000 or 2000 BNT is necessary.
And easy to implement as it is just a value change in the contract.
i disagree with the reduction from 2000 to 1000.
the cap is basically limiting the arb opportunity that we can handle and the incentive that is given to the caller. if the goal is to get more callers with higher incentive, the conversation should not be limiting this value but try to push it higher if anything.
1000 bnt at current rate is ~$400 gain.
to me, this seems like a very low cap and for no good reason.
my suggestion is to bring it back to 2000 and possibly even push it further up to 5000 bnt max cap.
If gas prices are too high, the fast lane halts and there are no arbs, and then it requires vote to increase it, which takes another few days while the fast lane isn’t functioning.
For that reason, I think it’s important to have a buffer while the system is being monitored and then we have some time to vote on a new value while the system is still running.
So I think 1000 gives us a good balance for the time being, not too low and not too high.
That’s correct, and I believe 1000 is actually large enough buffer that it won’t require frequent changes, but also not too high - I think that’s a good balance between the gas cost, frequency of updates and the caller profit.