Yes, if I get a bot I want to run seriously, i will use a vps, cheaper than spending £400 on a new laptop..
Trading What I see !?
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
If you're collecting data then a min spec server for about $20 a month is cheaper than the electric bill. And your laptop is usable again instead of being tied up for 21hrs a day.
Think I am going to give up trying to find a profitable bot for UK racing, it just seems far too efficient, "opportunities" of value seem very small, and get lost in hedging. I'm going to record more AUS race data, and see where it takes me.....
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
'Efficiency' is simply the fact that prices are generally 'correct' at anytime assuming you back/lay everything. Ie there's no magic moment when there's a bias vs bsp at scale. But prices move huge amounts throughout the day right up to the off which means that at any time and on any selection there's value to be had. It's 'just' a matter of comming up with yor own fair pricing algo to identify them.
UK racing is too liquid to just ignore, if you can't make value botting work (and that's tough to prove with so few races) then there's loads of niche areas where there's a dollar to be had. I mentioned price action around traditional odds values, behaviour after spikes, overbacking, MA crossovers etc etc, all worth thoroughly investigating and the more markets you watch the more potential niches you see.
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
There's no such thing as 'luck', only probablity And the more you bet the more outliers you'll see.
I don't watch the racing I'd imagine the rapidly changing weather is throwing up some unexpected results, not every horse likes a face full of driving rain and 10mins later the sun is shining. I know that feeling, I've had my wellies on and off and the kettle on and off all day and the new shed is still no more than a pile of timber. I've been poor value today.
So fair pricing algo, how can you do that on price action alone? Surely need fundamentals...ShaunWhite wrote: ↑Tue May 11, 2021 5:41 pm'Efficiency' is simply the fact that prices are generally 'correct' at anytime assuming you back/lay everything. Ie there's no magic moment when there's a bias vs bsp at scale. But prices move huge amounts throughout the day right up to the off which means that at any time and on any selection there's value to be had. It's 'just' a matter of comming up with yor own fair pricing algo to identify them.
UK racing is too liquid to just ignore, if you can't make value botting work (and that's tough to prove with so few races) then there's loads of niche areas where there's a dollar to be had. I mentioned price action around traditional odds values, behaviour after spikes, overbacking, MA crossovers etc etc, all worth thoroughly investigating and the more markets you watch the more potential niches you see.
Feeling frustrated again, partly because lost another load of money this week..
So been having some interesting discussions here and on Slack recently, and one common useful theme that comes out for Automation, is if you can't "Reason" your method/rule it's probably -EV. The bot I spent ages refining based on the "cutting losers" method, basically I started with a basic VWAP mean reversion value bot and found places where it was losing identifying the "pattern" and adding it to the Conditions of the bot. What I ended up with is something I believe is fitted to those "historic" patterns. Here are the Conditions for one of the rules of my bot, for you to all laugh at:
So I think the "Reasoning" is very important, not the actual "cutting losers".If your not enjoying it mate then your best to walk away. I know the feeling where it needs to be solved but sometimes you just have to say enough is enough. On the other hand if you want to continue, I’d take stakes down to nothing. Chucking money at the market will just eat it and add to the stress. Personally it sounds like your working really hard, probably too hard getting over complicated with it. Maybe go back to something real basic and let it run. You all sound much brighter then me but yea just a thought.
- Realrocknrolla
- Posts: 1903
- Joined: Fri Jun 05, 2020 7:15 pm