Good filter gone bad!
-
Im just after a little bit of advice regarding a filter which has had a big drop off in results over the last 3 months!
One of my SH goal filters has a very decent SR of post 60' goals. After removing certain leagues its running at 81% over 750 games, from last July through to present. However, there has been a big drop-off in results since July this year (nothing really appeared in June so July was purely summer leagues). The SR has dropped a lot down to less than 70%, albeit over only 80 or so selections. No matter what way i look at it (certain scores, leagues, using IP stats or pre-match only, tightening up criteria etc) the figures just dont look to improve. I have a second filter which has stood up ok through summer. I am probably 2-3 weeks away from seeing winter league selections appear
I was curious in this situation how everyone would advise playing the filter. On one hand i am considering stopping trading the filter until winter leagues have kicked in but on the other hand im conscious of missing a 'bounce back' in results, ie id be suffering the downside of the filter but then missing the upswing which often sees runs of 15-20 back to back wins. I know this sounds like FOMO but its more about riding out the variance. Im quite aggressive in staking, using 5% to back, so another option is to drop this to say 1-2% temporarily
Id be interested to hear thoughts on it
-
@lee-woodman First things is the staking for me, that has been picked up by others as well
Second hasn’t been picked up, when in this situation before I’ve set myself a deadline on it this could be timing or money it loses - then it goes into a data phase, because variance happens but we can tell ourselves to stick with it and its just variance - hence why I do if it loses X I’ll move it to tracking only.
If it comes back and proves itself it can go back in with money - if not you’ve cut the losses but still got the data.
-
@darri Appreciate the advice! I think you're right about the staking, there is probably no need for it to be as high as 5% right now as its more about the longer term for me. If i do 2% and make reasonable profit whilst gaining a lot more data then that cant be seen as anything but a success but if i make further losses then it wont be as damaging as 5% staking. I will stop this filter but continue to track as in time i will have decent numbers to draw decent conclusions from.
I guess i was trying to establish the fine line between 'persisting through the bad variance' and 'cutting losses and taking a step back' so yours and @Stuart-Capstick advice has helped a lot
-
@lee-woodman I think you already answered your own question and are simply asking am i correct to do those 2 things.
So yes wait id actually say stop trading this filter. If it is summer leagues that are not working now then leave them alone. It will only be a couple weeks until we have winter leagues in which teams have played at least either 5 games at home or 5 away which will then be able to pick them on the stats software filters. Use the rest of the summer leagues fixtures this year to test thoroughly. If they bounce back great you know for years to come that this is a good filter. If it continues to be sub par then you know youll have protected the bank. You dont want to keep adding in money to something that you have no confidence in or that is showing to be rather poor over the course of several months. Dont panic about a small time away tracking the data, your a good analytical trader with your head screwed on you know fine well all these decisions are just about planning for long term. 1 year of avoiding summer leagues could set you up for rest of trading career.
Yes your staking is too much for something your not confident about. Imagine you had £100k. Your telling people that your not confident in a trade but your past data says do it. You just wouldnt do it. Thats 5k! So dont do it now. 5% of a bank is a significant amount, you need to be super confident in a trade. Remember dont be greedy or try chase to a big bank. You could just bring these stakes down and then after every good year increase it my 1% upto 5%.
So yes wait, track summer leagues this year, dont trade them anymore this year, reduce staking until this is future proof. So rest of this year do 2% staking. If you come away with a solid year, then up them. But id be encouraging you to think about staking. 5% you need to be extremely confident that those trades are solid. Given the fact a downturn has happened might be time to reduce asap and rebuild confidence in it as it might not be as solid as you thought.
-
@stuart-capstick Post 70' has seen a similar drop-off, it was running at circa 68% but since July has dropped to about 58%. Similarly post 80' was circa 50% and has dropped to 42%. The least drop-off has been in a SHG which was just over 90% and is now about 87.5%. Obviously its a relatively small sample over 3 months so the further you look into data for patterns, the smaller the data set and therefore the less reliable.
My other filter has held up well and theres no overwhelming need to trade this so i think ill hold off trading it for now, continue to collect the data and see how it pans out
-
@lee-woodman
Hi Lee. As you say, the filter was running at 81% from 750 games. It doesn't look like a fluke.
All things being equal, there will likely be a return to that percentage. But from 750 games it would be very difficult to detect the patterns. Mathematically, you can detect likely swings if you have a large set of data (couple of thousand games).
Sounds like you have two very sensible options:- Wait
- Reduce stakes.
The good news is that you are approaching this perfectly. You aren't panicking or being affected by FOMO, but are looking at it rationally.
Just an idea: how does your filter hold up with goals after 70 mins? I ask because if backing the reward/risk ratio is better and most 2nd half goals come after 70 mins
