How much resources someone to operate such a bot puts into it, is likely dependent on profitability (operational cost vs. snatched sats). "Immense database" is a bit vague and I haven't thought about how much space an individual monitored entry would need. But profitability has very likely fallen significantly. Before the era of FullRBF the fastest bot had an edge (see screenshots at the end of my post). With FullRBF different bots turn sats "on" vulnerable private keys mostly into transaction fee where only the confirming mining pool profits (it would make sense for mining pools to secretly operate such bots). I can see a clear difference between those eras with an Electrum wallet I have (out of curiosity) where I put a few publicly exposed private keys into. (It's necessary to look into the transaction details, of course.) It still amazes me how often certain addresses are used for whatever purpose. Do I need to mention: don't ever use publicly available or derivable data for private keys and funds you want to keep!? If we had a reliable RBF history for transactions it could be possible to guesstimate the number of bots as long as individual bots would use only a single target address during their RBF fights. Maybe I setup a Core wallet for a few of the most used vulnerable private keys where I could watch the RBF history from Core's debug.log file.     Almost all sats gone to fees... I've no idea why e.g. someone sent 90,000,000 sats on 2022-07-27 to the "empty string" brainwallet. Apparently there were no RBF fights back then...
How much resources someone to operate such a bot puts into it, is likely dependent on profitability (operational cost vs. snatched sats). "Immense database" is a bit vague and I haven't thought about how much space an individual monitored entry would need. But profitability has very likely fallen significantly. Before the era of FullRBF the fastest bot had an edge (see screenshots at the end of my post). With FullRBF different bots turn sats "on" vulnerable private keys mostly into transaction fee where only the confirming mining pool profits (it would make sense for mining pools to secretly operate such bots). I can see a clear difference between those eras with an Electrum wallet I have (out of curiosity) where I put a few publicly exposed private keys into. (It's necessary to look into the transaction details, of course.) It still amazes me how often certain addresses are used for whatever purpose. Do I need to mention: don't ever use publicly available or derivable data for private keys and funds you want to keep!? If we had a reliable RBF history for transactions it could be possible to guesstimate the number of bots as long as individual bots would use only a single target address during their RBF fights. Maybe I setup a Core wallet for a few of the most used vulnerable private keys where I could watch the RBF history from Core's debug.log file.     Almost all sats gone to fees... I've no idea why e.g. someone sent 90,000,000 sats on 2022-07-27 to the "empty string" brainwallet. Apparently there were no RBF fights back then...

How much resources someone to operate such a bot puts into it, is likely dependent on profitability (operational cost vs. snatched sats). "Immense database" is a bit vague and I haven't thought about how much space an individual monitored entry would need. But profitability has very likely fallen significantly. Before the era of FullRBF the fastest bot had an edge (see screenshots at the end of my post). With FullRBF different bots turn sats "on" vulnerable private keys mostly into transaction fee where only the confirming mining pool profits (it would make sense for mining pools to secretly operate such bots). I can see a clear difference between those eras with an Electrum wallet I have (out of curiosity) where I put a few publicly exposed private keys into. (It's necessary to look into the transaction details, of course.) It still amazes me how often certain addresses are used for whatever purpose. Do I need to mention: don't ever use publicly available or derivable data for private keys and funds you want to keep!? If we had a reliable RBF history for transactions it could be possible to guesstimate the number of bots as long as individual bots would use only a single target address during their RBF fights. Maybe I setup a Core wallet for a few of the most used vulnerable private keys where I could watch the RBF history from Core's debug.log file.     Almost all sats gone to fees... I've no idea why e.g. someone sent 90,000,000 sats on 2022-07-27 to the "empty string" brainwallet. Apparently there were no RBF fights back then...

How much resources someone to operate such a bot puts into it, is likely dependent on profitability (operational cost vs. snatched sats). "Immense database" is a bit vague and I haven't thought about how much space an individual monitored entry would need. But profitability has very likely fallen significantly. Before the era of FullRBF the fastest bot had an edge (see screenshots at the end of my post). With FullRBF different bots turn sats "on" vulnerable private keys mostly into transaction fee where only the confirming mining pool profits (it would make sense for mining pools to secretly operate such bots). I can see a clear difference between those eras with an Electrum wallet I have (out of curiosity) where I put a few publicly exposed private keys into. (It's necessary to look into the transaction details, of course.) It still amazes me how often certain addresses are used for whatever purpose. Do I need to mention: don't ever use publicly available or derivable data for private keys and funds you want to keep!? If we had a reliable RBF history for transactions it could be possible to guesstimate the number of bots as long as individual bots would use only a single target address during their RBF fights. Maybe I setup a Core wallet for a few of the most used vulnerable private keys where I could watch the RBF history from Core's debug.log file.     Almost all sats gone to fees... I've no idea why e.g. someone sent 90,000,000 sats on 2022-07-27 to the "empty string" brainwallet. Apparently there were no RBF fights back then...