Afternoon dogs are a graveyard for some reason

[Fugazi]

Does it? That looks completely random.

Human behaviour.

1600-1700 people are leaving work, (yes people bet at work lol) picking up kids/taking them to hobbies. So bots take over. Then theyre back in the seat at 1700, and winding down/ doing other things later on.

Need a much bigger sample size to confirm my hypothesis.

Edit:

Alternatively, make 1600 positive. Pretty clear pattern then. Though my hypothesis wouldn't tie in with volumes I guess as late on volumes are good.

Guess I just come back when I have a sample of 60,000

I strongly recommend not going down this route, instead take a step back and try and understand the market better.

You want to quantify this behaviour into a signal, for example are the markets different morning/afternoon/evening due to liquidity / volatility / matching / amounts available / back or lay matching etc. If you can build a signal around this you will begin to understand the markets and then your own signals better and is very likely to take you down more 'fruitful' routes. Blindly dismissing markets due to time of day is not how you should be approaching betting.

Thats really good advice. Theres something impacting my automation - volume / volatility etc. My current approach of avoiding some parts of the day would just scrape some more pennies up. However if I understand what is actually impacting the automation, not only does this automation then have a much high ceiling, but as you said I'll start to really understand markets

[ShaunWhite]

Doesn't need me to agree but if you don't identify the material difference you'll miss winning markets in otherwise losing evenings and play losers during the otherwise profitable daytime. Not all evening races are different to all daytime races.

If you can define a 'market' in terms of its fundamental (aka meaningful) characteristics, then even the sport type can become a redundant variable.

[Fugazi]

Doesn't need me to agree but if you don't identify the material difference you'll miss winning markets in otherwise losing evenings and play losers during the otherwise profitable daytime. Not all evening races are different to all daytime races.

If you can define a 'market' in terms of its fundamental (aka meaningful) characteristics, then even the sport type can become a redundant variable.

Been using BA around 15 months now. Profitable for maybe the last 4, but been stuck in a rut of tweaking automations in an attempt to scale and reducing my P and L to peanuts doing so.

The penny has finally dropped I need to spend the time setting up and storing data properly. Though thats easier to spot now I have an automation I know works but needs optimising. Much more motivated now I know storing data is going to increase my profits for sure.

[ShaunWhite]

been stuck in a rut of tweaking automations in an attempt to scale and reducing my P and L to peanuts doing so.

#1 sample size, it's impractical to run it long enough to base a change on.
#2 when you do make a change it's running on different markets so you don't know if your change was effective or if something in the markets that's changed.

You're kind of hamstrung because you can't run (or test) several variations on the same markets.

In meantime download and dig into that bet history, you should be able to see what prices, timeframe, stake and side is making or losing money. Knock up a template spreadsheet and do it every day or week.