Re : Does Scalping work anymore

A place to discuss anything.
Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 3:43 pm

I have been very sceptical about one tick scalping for some time.

I understand the theory.

On average, the true price lies somewhere between the spread. Yes, the current price may not be the true price, but over thousands of bets, for every time you're x% above the true price, you'll be x% below the true price, and it will even itself out.

So the principle is that, by offering to the market, you beat the spread and obtain long term backing and laying value (and thereby obtain long term trading profitability).

However, my previous experiences of it have been pretty bad.

I'd long since dismissed the approach, but the spreadsheet in the locked thread on this matter - viewtopic.php?f=2&t=159&start=20 - gave me pause for thought. Those results are very impressive (as are the ones in this thread).

Perhaps I had been too hasty. And perhaps I hadn't used a large enough sample previously, or didn't operate in a sufficiently mechanical way. So I created a simple spreadsheet today that does the following:

- £2 offers on both sides of the book.

- If an offer is taken, I make another offer on the same side of the book.

- If the price moves, repeat the process at the new price (without having more than one offer on either side of the book at any time).

- Green or red up 10 seconds before the off.

In theory, this should work great. Win and little, lose a little, secure small but significant long term edge.

However, using £2 stakes, I lost £7 in my first race and £5 in my second.

Maybe I was just extremely unlucky, maybe the market has changed, or maybe you need a faster internet connection and better computer than I have (I use 0.2 refreshes as my computer doesn't cope well with full stream). Or yes, maybe my approach was flawed.

However, with those kind of losses, which far exceed any losses in the spreadsheet, I'm not going to find out. :)

I hope the above doesn't come across as provocative. I know there are people here who swear by scalping, and I'd be genuinely interested in hearing people's opinions on my experiment.

Jeff

PS LeTiss - before you ask - the answer is 'No'. :lol:

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 4:05 pm

PPS Stats for the 2 races in question:

297 bets placed.

£771.71 turnover

£12.68 loss

-1.64% ROI

User avatar
Euler
Posts: 14567
Joined: Wed Nov 10, 2010 1:39 pm
Location: Bet Angel HQ
Contact:

Thu Oct 12, 2017 4:21 pm

What were the two races?

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 4:23 pm

GB / Ayr 12th Oct / 15:05 1m Hcap

GB / Worc 12th Oct / 15:25 2m Stks NHF

User avatar
Euler
Posts: 14567
Joined: Wed Nov 10, 2010 1:39 pm
Location: Bet Angel HQ
Contact:

Thu Oct 12, 2017 4:25 pm

Fav?

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 4:33 pm

Yes.

When the favourites switched, I ended up trading both favourites, as I hadn't programmed by spreadsheet to stick to the horse that was favourite when I arrived at the market.

I appreciate that higher priced horses are generally better for scalping, but as you had traded low priced horses in your 'random scalping' experiment, I thought I'd include them and see what happened.

I've attached my results.
Scalping Experiment.xls
Jeff
You do not have the required permissions to view the files attached to this post.

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 4:42 pm

Maybe two races aren't representative.

Plus, the fact that the bot stopped trading a favourite when it stopped being a favourite may not have helped (I've since amended it so that it only trades the horse that's favourite when I arrive at the market).

I'll trade the rest of today's races using 10p stakes and let you guys know how I get on.

Jeff

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 5:59 pm

I've decided to call time on my experiment.

Since my post below about an hour ago, the results are as follows:

Stake: 10p
Number of bets: 1073
Turnover: £140.54
P/L: £2.94 loss
ROI: -2.09%

I've excluded races where, due to a bug in my bot, I traded more than one horse.

Possible reasons why I wasn't more successful:

- I was trading the favourite.
- I was using a 0.2 second refresh

Any thoughts?

I've attached my results spreadsheet.

Jeff
Scalping Experiment II.xls
Ferru123 wrote:
Thu Oct 12, 2017 4:42 pm
Maybe two races aren't representative.

Plus, the fact that the bot stopped trading a favourite when it stopped being a favourite may not have helped (I've since amended it so that it only trades the horse that's favourite when I arrive at the market).

I'll trade the rest of today's races using 10p stakes and let you guys know how I get on.

Jeff
You do not have the required permissions to view the files attached to this post.

User avatar
LeTiss
Posts: 3881
Joined: Fri May 08, 2009 6:04 pm

Thu Oct 12, 2017 6:20 pm

I don't know if you saw Peter's post the other day on automation on Greyhounds.

He had a cracking day, bagging £144 just on dog automation, but Peter stressed that it was a bot that was on it's 27th version.

Don't scrap everything Jeff, just try to separate the aspects that are good from the bits that are shit, and then try to re-work it using the successful elements

Ferru123
Posts: 6700
Joined: Fri Dec 11, 2009 10:51 pm

Thu Oct 12, 2017 6:28 pm

Thanks LeTiss

I agree. :)

Today's experiment was to test 'unrefined' blind scalping, but the result doesn't mean that scalping can't be made to work.

If I throw some directional bias or wom into the mix, the result may be different.

Jeff

Post Reply
  • Information
  • Who is online

    Users browsing this forum: AlanArneill, Baidu [Spider] and 3 guests