Trading at random
Hey Peter,
I have no problem winning when I stick to gambling rather than this trading mumbo-jumbo. I will NOT be wasting my time with conventional trading (i.e., use of stop-loss with a predefined exit strategy) again, there will be no more stop-loss nonsense for me.
Make no mistake: Eventually I will return to this game in a big way, move in with big stakes, and personally crush every single one of the traders.
Cheers.
I have no problem winning when I stick to gambling rather than this trading mumbo-jumbo. I will NOT be wasting my time with conventional trading (i.e., use of stop-loss with a predefined exit strategy) again, there will be no more stop-loss nonsense for me.
Make no mistake: Eventually I will return to this game in a big way, move in with big stakes, and personally crush every single one of the traders.
Cheers.
Thanks for sharing this Peterle.. its very generous of you..great work.PeterLe wrote:Not sure if anyone will find this interesting, but I'll share it anyway..
From time to time I like to check the market and I wanted to test a theory - I ran a scalping test today on a spare account.
30 markets
591 bets
Entry Time : 180 Secs
Total Stakes: £5,648
The stakes were variable based on liability
P&L: + £8.12 (after commission)
Unmatched bets converted to take SP 3 seconds before the off
9 races were <0, 21 Positive (4 out of 6 races at Ballin were negative, very volatile this evening and someone was trying to manipulate the prices with 5 figure stakes!
Over 90+% of the losses came from just three races (so in theory, it would have been possible to manually manage these)
It ran unaided and I checked it from time to time to check it was going well
regards
Peter
-
spreadbetting
- Posts: 3140
- Joined: Sun Jan 31, 2010 8:06 pm
The thing is I'm sure Peter is very accomplished at writing bots so it's very unlikely his idea of 'random' will be the same as another persons take on 'random' especially if they already have a view on how markets work.
Hispreadbetting wrote:The thing is I'm sure Peter is very accomplished at writing bots so it's very unlikely his idea of 'random' will be the same as another persons take on 'random' especially if they already have a view on how markets work.
That's a fair point.
The BOT was really looking for one tick scalps and I did apply some very basic logic to determine when to fire.
What I have found is that you can compile very complex spreadsheets that look for the perfect moment and fire in a bet. Although you have a high success rate you tend not to put as much through the market. (if you use very large stakes, you appear on the radar and it makes getting matched harder). I think its better to fire in lots of smaller bets as it spreads your risk a little too.
This test was used on every market; even markets were the fav was less than 2's (ie Markets that are not ideal for scalping in the first place), so just removing them alone would increase the strike rate
Hope that helps
Regards
Peter
-
steven1976
- Posts: 1744
- Joined: Tue Jan 19, 2010 6:28 am
Do you run it on just the favorite or have you tried it across the field?
I was crunching some numbers and thought I would qualify my comments on random trading by looking at some recent data.
I looked at 6598 races. To 'unbias' the entry and exit I took the mid price of the current open and closing prices. If you were foolish you could make the results slightly worse and if you were smart slightly better. That's up to you!
5% of races broke even and the others were close to 50% but will a slight bias. If we punish ourselves by taking the worst possible result from this bias. You would have ended up losing 8p on average per race using £10 stakes over the period.
So the conclusion is, there is a very small barrier to overcome to become profitable. Which means the situation hasn't changed much from when I first looked at market all those years ago.
I'm still very confident that you can overcome this just by making slightly better decisions than random, or other, less educated market participants.
The state of trading education in general is such that I think a lot of the duff advice out there probably creates an opportunity in itself for people who are prepared to work hard to overcome this small barrier you see in the markets. You only need to be slightly better than other participants to make money.
I looked at 6598 races. To 'unbias' the entry and exit I took the mid price of the current open and closing prices. If you were foolish you could make the results slightly worse and if you were smart slightly better. That's up to you!
5% of races broke even and the others were close to 50% but will a slight bias. If we punish ourselves by taking the worst possible result from this bias. You would have ended up losing 8p on average per race using £10 stakes over the period.
So the conclusion is, there is a very small barrier to overcome to become profitable. Which means the situation hasn't changed much from when I first looked at market all those years ago.
I'm still very confident that you can overcome this just by making slightly better decisions than random, or other, less educated market participants.
The state of trading education in general is such that I think a lot of the duff advice out there probably creates an opportunity in itself for people who are prepared to work hard to overcome this small barrier you see in the markets. You only need to be slightly better than other participants to make money.
If performance is measured relative to the market, it does indeed appear that there is a very small barrier to profitability. But if we compare performance relative to other participants, an Eeyore might come to a more pessimistic, but equally valid, conclusion.
If 5% of participants are consistently successful (perhaps an overestimate) then someone sitting at 50% in the competence rankings needs to become 45% "better" in order to reach profitability. Someone languishing at 5% needs to boost their performance by a factor of 19.
Looked at in Peter's way, the aspiring trader needs to drive over a speed hump. From an Eeyore perspective, it is more like scaling the north face of the Eiger.
And they are both correct.
If 5% of participants are consistently successful (perhaps an overestimate) then someone sitting at 50% in the competence rankings needs to become 45% "better" in order to reach profitability. Someone languishing at 5% needs to boost their performance by a factor of 19.
Looked at in Peter's way, the aspiring trader needs to drive over a speed hump. From an Eeyore perspective, it is more like scaling the north face of the Eiger.
And they are both correct.
I think the 5% figure for successful traders is correct. Taking the figure of 1 in 200 Betfair accounts long-term profitable, and excluding the 'inactive' accounts (90%), that shows 1 in 20 (5%) successful. Also this is supported by Google research, which shows a failure rate for traders in general (across all financial markets) in the ballpark of 95%, with 5% successful long-term.Akindwurd wrote:
If 5% of participants are consistently successful (perhaps an overestimate) then someone sitting at 50% in the competence rankings needs to become 45% "better" in order to reach profitability. Someone languishing at 5% needs to boost their performance by a factor of 19.
The 1 in 200 figure for the long-term successful Betfair accounts has been stated by Betfair management. It seems reasonable to exclude the inactive accounts, that would imply 5% are successful.
If you Google then the estimates for the % long-term successful traders in financial markets seem to be in the 5-10% ballpark.
Sure it's educated guesswork, I reckon the 5% figure probably isn't far wrong on average?
If you Google then the estimates for the % long-term successful traders in financial markets seem to be in the 5-10% ballpark.
Sure it's educated guesswork, I reckon the 5% figure probably isn't far wrong on average?
"What percentage of your users make a profit each year?
From memory, c.25% of our customers are profitable after commission is deducted."
viewtopic.php?f=15&t=9338
From memory, c.25% of our customers are profitable after commission is deducted."
viewtopic.php?f=15&t=9338
'Profitability' should mean 'the % of accounts making a long-term profit', not 'the % of accounts in profit at a given time' - the latter measure is not accurate, because most accounts are 'inactive' (users only making a few bets then never using the account again).LinusP wrote:"What percentage of your users make a profit each year?
From memory, c.25% of our customers are profitable after commission is deducted."
viewtopic.php?f=15&t=9338
Only 1 in 200 accounts pay premium charge, so only in 1 in 200 are actually profitable long-term.