Spent all evening running new backtests and various simpler models of AUS data (minus NSW), with commission correctly processed!
This is my best effort for today, i'm knackered and only crap to show for it!
Trading What I see !?
I'm not sure i'm going to succeed with attempting to alter parameters for all the algo bits, WoM, levels, averages, vwap, money, prices, ... without "reason". A market is a random set of punters/participants/bots, which is probably why backtests show graphs with bits doing well, then bits where it does awful, i'm guessing these different phases are different combinations of "participants"... So how do you recognise participants with a bot?
I really can't be arsed to backtest some more variations until i've got an idea in my head of how to progress in a positive direction.
I'm just going to run my Betfair data leecher, not worth running a live bot that's just losing me money...
At least by the time I get something in my head i'll either have loads of data or a closed account!!
I'm just going to run my Betfair data leecher, not worth running a live bot that's just losing me money...
At least by the time I get something in my head i'll either have loads of data or a closed account!!
- ShaunWhite
- Posts: 9731
- Joined: Sat Sep 03, 2016 3:42 am
It's not random, it's the age old game of supply and demand.goat68 wrote: ↑Fri Jun 04, 2021 12:34 amA market is a random set of punters/participants/bots, which is probably why backtests show graphs with bits doing well, then bits where it does awful, i'm guessing these different phases are different combinations of "participants"... So how do you recognise participants with a bot?
If it's showing a random outcome then the characteristics you're watching aren't relevent. That's why you need to look at those PL distributions as you alter parameters to sort the wheat from the chaff.
Why have you made it so complicated!? You know you can't possibly understand the effect of so many conflicting metrics, so if it's not possible then it's the wrong approach. Not being able to think of the right approach isn't a reason to keep doing something you know is wrong.
It’s a daunting task to have so many bots up and running. I think it’s good to run one stat alone .goat68 wrote: ↑Fri Jun 04, 2021 12:37 amI really can't be arsed to backtest some more variations until i've got an idea in my head of how to progress in a positive direction.
I'm just going to run my Betfair data leecher, not worth running a live bot that's just losing me money...
At least by the time I get something in my head i'll either have loads of data or a closed account!!
If it’s BSP , or VWAP etc . I try to stick to one and run all sorts of variations .
Every loosing bot has a winning bot,
may be it’s good to try the opposing bot against your loosing one. If you bot is meant to lay , then switch it to back, and run the two variations at the same time . You might see another variance emerge and that’s what I call info.
I am only testing one bot / one stat and that’s it . Very slow step by step process , Especially when not backtesting .
NB : fill rates and human behaviour are hard to backtest and your bot might have had a profitable result , were it to hit a live market.
Why? Can short term volatility not be taken advantage of at a price that isn't value?LinusP wrote: ↑Thu Jun 03, 2021 2:40 pmNo you can'tmarko236 wrote: ↑Thu Jun 03, 2021 11:56 amI only use the word value when betting, a horse for example can be in a non value price but you can still make money on it trading. I don't care or even think about value as a trader i'm just looking for signals that the market is more likely to move in a certain direction.
isn't short term volatility "noise", which by definition is random, hence in the long term will be net 0 ?jamesg46 wrote: ↑Sat Jun 05, 2021 11:31 pmWhy? Can short term volatility not be taken advantage of at a price that isn't value?LinusP wrote: ↑Thu Jun 03, 2021 2:40 pmNo you can'tmarko236 wrote: ↑Thu Jun 03, 2021 11:56 amI only use the word value when betting, a horse for example can be in a non value price but you can still make money on it trading. I don't care or even think about value as a trader i'm just looking for signals that the market is more likely to move in a certain direction.
Not sure Goat, you tell me. Liam certainly knows what he's talking about & so does Peter, but....
https://youtu.be/i0tzkuqcwWQ
Tell me, is anyone seeking value in this trade?
- wearthefoxhat
- Posts: 3221
- Joined: Sun Feb 18, 2018 9:55 am
Not in the sense of a numerical amount that is a higher/lower than another, maybe more in the value of the strategy to an individual that it can make a profit long-term and the right judgement is used in choosing the right market/selection.jamesg46 wrote: ↑Sat Jun 05, 2021 11:42 pmNot sure Goat, you tell me. Liam certainly knows what he's talking about & so does Peter, but....
https://youtu.be/i0tzkuqcwWQ
Tell me, is anyone seeking value in this trade?
This video is definitely a good one for any trader that enjoys manual trading. I've been guilty in the past of wanting to find the holy grail of full automation, (settled on semi-automation) but this approach can be adapted for swing trades and increase profits longer term, manually.
not sure either, but guess there's more than one way to peel an orange!jamesg46 wrote: ↑Sat Jun 05, 2021 11:42 pmNot sure Goat, you tell me. Liam certainly knows what he's talking about & so does Peter, but....
https://youtu.be/i0tzkuqcwWQ
Tell me, is anyone seeking value in this trade?
Yeah, I think that's fair. There is definitely value in the effectiveness of using that strategy with short term volatility.wearthefoxhat wrote: ↑Sun Jun 06, 2021 7:11 amNot in the sense of a numerical amount that is a higher/lower than another, maybe more in the value of the strategy to an individual that it can make a profit long-term and the right judgement is used in choosing the right market/selection.jamesg46 wrote: ↑Sat Jun 05, 2021 11:42 pmNot sure Goat, you tell me. Liam certainly knows what he's talking about & so does Peter, but....
https://youtu.be/i0tzkuqcwWQ
Tell me, is anyone seeking value in this trade?
This video is definitely a good one for any trader that enjoys manual trading. I've been guilty in the past of wanting to find the holy grail of full automation, (settled on semi-automation) but this approach can be adapted for swing trades and increase profits longer term, manually.
If you're going to use that strategy would you be taking a value price every time, be it you know you are or not?. What exactly is value? Is it price, is it strategy, is it opportunity, or a mix of them all.
Is every price with matched volume value? From the beginning to the end of a markets time, someone somewhere is taking a price they feel is either a value entry or exit based on any number of emotions or action taking place. For anyone buying there has to be someone selling, are they both taking value?.
Is a value price a price within any given strategy that any one individual is using at any one time. What the hell is value
In order to be profitable, one must be taking value. If you are trading volatility it doesn't matter, your strike rate still needs to be better than the implied probabilities in the markets. To say you do not need to do "value taking" in order to be profitable is an oxymoron - profitability is an outcome of taking bets at value.
Sure - volatility traders may not be assessing value with the same methodology a low-frequency trader might. But the end of the day, when you load your PNL page, you are only going to get +ve pnl if your strike rate is better than the implied strike rate from pricing. And thus you are taking at value.
Successful order flow trading/volatility trading is still value betting by proxy. Any profitably betting operation is. Let's stop this bs "you don't need to take value if you are "trading""
So once again - it is NOT possible to be profitable UNLESS you take/place value bets.
Sure - volatility traders may not be assessing value with the same methodology a low-frequency trader might. But the end of the day, when you load your PNL page, you are only going to get +ve pnl if your strike rate is better than the implied strike rate from pricing. And thus you are taking at value.
Successful order flow trading/volatility trading is still value betting by proxy. Any profitably betting operation is. Let's stop this bs "you don't need to take value if you are "trading""
So once again - it is NOT possible to be profitable UNLESS you take/place value bets.