Automation Computational Efficiency?

[DrJAT]

This is more of a technical question really, rather than a trading one. I'm just wondering how computationally efficient the Bet Angel automation is? Is there any practical advantage to configuring my BAF files with as few rules as possible and eliminate "unnecessary" steps? A few of the kinds of things I'm thinking of are:

1. Making stored values for prices and things that I don't strictly need but help with debugging if and when things go slightly wrong.
2. Updating stored values more frequently than is strictly necessary.
3. Creating the same stored value multiple times - once just for use within the market and once for sharing with other markets at an event level. I know the latter would serve both purposes but I don't like sharing variables outside of their intended scope unless necessary and it makes it easier when duplicating rules. Like if I've got a bunch of drip back bet rules using prices from stored values but some of them are stored at market level and others at event level it's a recipe for creating a problem that could be a pain to spot, so it's easier to have them all set at market level and just create a separate value to share at event level if I need to.

How much computational power does it take checking conditions compared to actually firing a rule? For example, if I write a BAF with a bunch of rules that only ever fire when certain conditions are met and apply it to every football match on a Saturday afternoon, is this execution going to slow down my refresh by keep checking every game even though the rules rarely kick in?

Thanks in advance!

[Derek27]

This is more of a technical question really, rather than a trading one. I'm just wondering how computationally efficient the Bet Angel automation is? Is there any practical advantage to configuring my BAF files with as few rules as possible and eliminate "unnecessary" steps? A few of the kinds of things I'm thinking of are:

1. Making stored values for prices and things that I don't strictly need but help with debugging if and when things go slightly wrong.
2. Updating stored values more frequently than is strictly necessary.
3. Creating the same stored value multiple times - once just for use within the market and once for sharing with other markets at an event level. I know the latter would serve both purposes but I don't like sharing variables outside of their intended scope unless necessary and it makes it easier when duplicating rules. Like if I've got a bunch of drip back bet rules using prices from stored values but some of them are stored at market level and others at event level it's a recipe for creating a problem that could be a pain to spot, so it's easier to have them all set at market level and just create a separate value to share at event level if I need to.

How much computational power does it take checking conditions compared to actually firing a rule? For example, if I write a BAF with a bunch of rules that only ever fire when certain conditions are met and apply it to every football match on a Saturday afternoon, is this execution going to slow down my refresh by keep checking every game even though the rules rarely kick in?

Thanks in advance!

I would imagine the time taken to manipulate a stored value would be measured in microseconds. If you have a long list of stored value operations running on every refresh, refreshing every 20ms and on dozens of markets I wouldn't expect any performance issues on an average PC. The bottleneck is usually on the number of markets you're watching simultaneously.