Ok I have been paper trading a very simple system that I found online that has been making profit for the last 7 yrs. It has had 19 winners out of 42 selections and made a 23.8 point profit. The rules make sense and there are only 2 -3 bets per week usually, though recently there were 6 in one day, 4 lost one won at 6/1, the other at 8/11
Now what I want to know is are these amazing results likely to continue or is it not realistic to have any sort of big advantage over the bookies
Paper trading new system
Due to random chance, you will often get long winning periods and long losing periods (you can verify this using my simulator spreadsheet - viewtopic.php?f=2&t=1543&p=16088&hilit=simulator#p16088).
Therefore, although I'm no statistician, I'd say you need a much larger sample of bets to draw any conclusions about your system's long-term effectiveness.
Jeff
Therefore, although I'm no statistician, I'd say you need a much larger sample of bets to draw any conclusions about your system's long-term effectiveness.
Jeff
Lafaso870 wrote:Ok I have been paper trading a very simple system that I found online that has been making profit for the last 7 yrs. It has had 19 winners out of 42 selections and made a 23.8 point profit.
I wrote an article about paper trading recently. I'm not an advocate of paper trading as you have a tendancy to see things that just don't exist in reality or just backfit data. Also the process of executing a system often influences the results. I tried loads of paper systems for years before I realised the only way to make true progress is to be active in the market taking real risk.
Putting aside the issue of whether backtesting is a good idea, does anyone know how you would determine whether a set of results was statistically significant?
For example, let's say I bet on 20-1 horses and I show a 5% profit after 3,000 bets. How would I work out how likely it is that that's down to random chance?
Jeff
For example, let's say I bet on 20-1 horses and I show a 5% profit after 3,000 bets. How would I work out how likely it is that that's down to random chance?
Jeff
-
CaerMyrddin
- Posts: 1271
- Joined: Mon Sep 07, 2009 10:47 am
I won't be working the figures, but an educated guess would be around 30%?
when you are paper trading / back testing systems,
you need to take into you account the big losing runs you will inevitably have.
for example in real life would you still be throwing money at a system where you are losing money hand over fist if you didn't know it was going to end up in profit.
look a the lowest point of the system and if it's bad
be honest and ask yourself would you have give up at that point not knowing what happens in the future.
you need to take into you account the big losing runs you will inevitably have.
for example in real life would you still be throwing money at a system where you are losing money hand over fist if you didn't know it was going to end up in profit.
look a the lowest point of the system and if it's bad
be honest and ask yourself would you have give up at that point not knowing what happens in the future.
Last edited by freddy on Fri Nov 26, 2010 5:31 pm, edited 1 time in total.
Also years ago when i began messing about on the exchanges, i developed a laying system and when i was paper testing it i had a run of 142 losing lays in a row at under 10-1 odds 
.
but amazingly the system turned out to be a big loss maker
,
just goes to show how hard it is and how much data you need to prove it works.
i just trade now much much easier
.but amazingly the system turned out to be a big loss maker
, just goes to show how hard it is and how much data you need to prove it works.
i just trade now much much easier
