Losing strategies wanted
-
- Posts: 55
- Joined: Wed Sep 30, 2015 8:26 pm
ive been backing fallen horses now for the past week. yet to win a penny *taps nose*
it is.steven1976 wrote: ↑Tue Sep 16, 2014 1:31 pmTry it on a horse about 10s. Your thinking that as a result of a race it would break even over time. Put a 10 pound back or lay into the market and do nothing else other than sell out when the next move happens either against you or with you for 1 tick. You may be surprised how random, random really is.
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Been working on this throughout April (pre-off lay strat - following trends), decided to fully incorporate all the favourites, & I get this equity curve (400 trades)
Peter if you're interested I'll drop you a message
Peter if you're interested I'll drop you a message
What does it look like if you flip it?ruthlessimon wrote: ↑Mon Apr 30, 2018 5:13 pmBeen working on this throughout April (pre-off lay strat - following trends), decided to fully incorporate all the favourites, & I get this equity curve (400 trades)
Peter if you're interested I'll drop you a message
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Swap the lay for a back, there aren't stops & targets hence why it's completely inverted - I personally, along with another member have doubts as to whether this is an inherent bias or just the byproduct of randomness
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Just noticed the scale is off, gonna quickly find out why that's the case
But you want to give it away?ruthlessimon wrote: ↑Mon Apr 30, 2018 5:47 pmSwap the lay for a back, there aren't stops & targets hence why it's completely inverted - I personally, along with another member have doubts as to whether this is an inherent bias or just the byproduct of randomness
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
I'd bet Peter already knows; & the problem is this pattern is consistent on data where it shouldn't.
Changing the variables don't affect the equity curve. I've got 30 variations of this strategy, all they're all negative (each roughly the same amount) - & the ones I hoped would be zero expectancy aren't.
Hence why I think wtf is going on with this!!
(btw the scaling error was 97, not 397 )
Changing the variables don't affect the equity curve. I've got 30 variations of this strategy, all they're all negative (each roughly the same amount) - & the ones I hoped would be zero expectancy aren't.
Hence why I think wtf is going on with this!!
(btw the scaling error was 97, not 397 )
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
For 10 iterations (delayed entry), in a sheer profitability sense, they all fall roughly around -200ticks, hence why there are serious questions - iteration 0 (a blind time-based lay) shouldn't be that negative imo
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Just got today's afternoon data (not used in the above graphs). 5 trades, 3 losers, 2 scratches - 15:40 got marked as a large loser -13ticks
(iteration 8)
(iteration 8)
-
- Posts: 1084
- Joined: Fri Nov 20, 2015 9:38 am
Is it as simple as you are "positively selecting" the wrong markets for strategy due to entry criteria?
I've got a similar problem I'm trying to figure out and that's my best guess atm - still working on it...
I've got a similar problem I'm trying to figure out and that's my best guess atm - still working on it...
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Just to clarify what do you mean by "positive selection"?
-
- Posts: 1084
- Joined: Fri Nov 20, 2015 9:38 am
Say you are laying 5 ticks under market price and get matched, then (in general) you are more likely to be matched by by selections that come in and then keep on coming in (as opposed to heading back out)... i.e. you are being "positively" selected against.... just a thought - I think something like this is happening in one of my strategies...
In assurance terms : a sick person is more likely to buy sickness insurance...
Also, are you offering money or taking price on offer - what would a 1 tick difference at entry do for your results?
In assurance terms : a sick person is more likely to buy sickness insurance...
Also, are you offering money or taking price on offer - what would a 1 tick difference at entry do for your results?
- ruthlessimon
- Posts: 2095
- Joined: Wed Mar 23, 2016 3:54 pm
Ah I see
No this is not the case in the above strategies; & in fact, it's the complete opposite. Which is why I expected these strategies to be profitable - hence why I thought they'd be great for this thread
The iterations above are 1 tick differences - i.e. Critera met, + 1 drift tick = Iteration 1, Critera met + 5 drift tick = Iteration 5 etc etc
My data is streamlined to LTP so I'm unsure - i.e.
No this is not the case in the above strategies; & in fact, it's the complete opposite. Which is why I expected these strategies to be profitable - hence why I thought they'd be great for this thread
The iterations above are 1 tick differences - i.e. Critera met, + 1 drift tick = Iteration 1, Critera met + 5 drift tick = Iteration 5 etc etc
My data is streamlined to LTP so I'm unsure - i.e.
I have often thought that doing the opposite to a person losing would work but there are a lot of factors. the main one is getting matched in the first place. Im not an expert but I think the main money in this game is knowing how to spot the bad trades by newbies. The people that try Betfair for the first time and have just done a one-off random bet, especially.