Steamers & Drifters

[ruthlessimon]

& here's why it's tricky.

caveat, this is an extremely basic strategy, & this is only using April's markets:



A trader punting a lay if the fav drifted 15 ticks lost less money, than the trader punting a lay if the fav drifted 5 ticks.

& the graph gives us a clue as to why.

Frequency. There were simply way less markets that drifted 15 ticks. & when they did drift 15 ticks the entry time was so late, there was no time to damage that trader with a powerful reversal

So joining a trend early isn't necessarily good; & my data suggests it's quite a bit worse (-250ticks vs -120ticks).

[firlandsfarm]

I don't think there is actually a dictionary def of timescales and amounts, it's more of an adjective than a noun.

I appreciate that Shaun, that's why I thought it might be interesting to see what timescales and amounts people do apply.

[firlandsfarm]

Simon are they graphs for laying the Fav when it's price drifts by 5/15 ticks and holding until the race is run? The line showing the loss incurred? Interesting ... is the loss too small per race to make taking a contrary position profitable and in the end commission is the only winner!

[Derek27]

If a horse is backed from 8 down to 5, does it really matter whether or not it's classed as a steamer?

Numbers speak louder than words!

[firlandsfarm]

If a horse is backed from 8 down to 5, does it really matter whether or not it's classed as a steamer?

Not at all and I never suggested it mattered, I was just interested in what in people's mind makes a horse a Steamer or Drifter, it's a purely academic question ... same as if a meal is big or small does it really matter if it's classed as big or small so long as it satisfies? But if someone comments on a restaurant and says the portions are big or small you tend to have an expectation.

[ruthlessimon]

Simon are they graphs for laying the Fav when it's price drifts by 5/15 ticks and holding until the race is run? The line showing the loss incurred? Interesting ... is the loss too small per race to make taking a contrary position profitable and in the end commission is the only winner!

No, they're all pre-off. The end of each trade is exactly 00:00:00.

That's correct it's showing the cumulative loss had you performed the strategy throughout April.

Good question. It's hard to gauge - although I'm inclined to say no, the trade reversed can work:



This is the 5tick entry graph, with the trades in order from "worst to best". What's interesting is visually far more are over than -10 threshold than above the +10 threshold.

Overlaying the winners to losers they seem to be outperforming across every aspect. Not only are they more frequent, but you lose more each time

[firlandsfarm]

No, they're all pre-off. The end of each trade is exactly 00:00:00.

OK, sorry Simon but for clarity ... are you saying monitor price over "T" minutes before the off and if the price moves out by 5 ticks at any time within the T minutes then open a trade and close at BSP?

I'm inclined to say ... the trade reversed can work

Now you are doing this on laying a 5 point drift so "worse" is good if backing and therefore there are more +10 backers than -10.

I have always wondered if Steamers and Drifters were viewed the wrong way round and this was in the back of my mind when I asked the question. I'm thinking back a Drifter because the odds are likely to go out too far and lay a Steamer because they come in too far. Same as financials, they always run too far and have a correction. The difference here though is because race betting is finite with a closing deadline they may not always get the chance to be corrected.

Time for me to start collecting pre-off data I think, I had better go and find that spreadsheet.

[ruthlessimon]

OK, sorry Simon but for clarity ... are you saying monitor price over "T" minutes before the off and if the price moves out by 5 ticks at any time within the T minutes then open a trade and close at BSP?

Yes, that's how the experiment was setup.

Data capture begins at 5mins. The price of the fav @ 5mins is recorded & remembered. If the fav drifts 5 ticks from this price, lay it. Exit the position at 0mins. The problem is, every variation (no ticks, 5 ticks, 10 ticks, 20 ticks) came out negative - therefore the data suggests if blindly trading by price, never lay the fav. There's something amiss.

Now you are doing this on laying a 5 point drift so "worse" is good if backing and therefore there are more +10 backers than -10.

That's right & that's what the data suggests. If you lay a 5tick drift, it's more common that it will actually steam 10ticks after your entry.

I have always wondered if Steamers and Drifters were viewed the wrong way round and this was in the back of my mind when I asked the question. I'm thinking back a Drifter because the odds are likely to go out too far and lay a Steamer because they come in too far. Same as financials, they always run too far and have a correction. The difference here though is because race betting is finite with a closing deadline they may not always get the chance to be corrected.

Time for me to start collecting pre-off data I think, I had better go and find that spreadsheet.

Now you've hit a really interesting crux, & I mentioned this in a previous post. Trading a steamer/drifter, is totally different to labelling a steamer/drifter. This is where "strength & weakness" enter the game. A "drifter" isn't necessarily "weak" (& that's what the above data suggests). Maybe another thread labeled "Strength & Weakness" is needed. however, I doubt that thread will get replies publicly.

For instance, take the above trades. The hypothesis of the trade idea was: If the fav drifts 5 ticks, it should drift further - that hypothesis simplified to two words - (drifter, weakness). "Drifter" past tense, "weakness", future tense.

[napshnap]

ruthlessimon

How do you define in your calculations that runner is a fav? If its price is 1 tick lower than 2nd favs price - is it still a fav?

that hypothesis simplified to two words - (drifter, weakness).

Do you mean drift trend weakness?

[firlandsfarm]

[The hypothesis of the trade idea was: If the fav drifts 5 ticks, it should drift further

Simon, thanks for your input into this, subjects need to be explored to be advanced. I wasn't actually thinking it "should drift further", I was thinking it now offers more (back) value making it more attractive if it doesn't correct back in or if it does correct then it is tradeable so a possible mix between a trade and if that fails then a value bet (ignoring an in-running trade out which brings another variable into the mix). Your research is sufficient to make me take it further, but first I need to gather the data, all my data is about the runners and riders ... I have no trading data to delve into. Again, thanks for your help.

[ruthlessimon]

How do you define in your calculations that runner is a fav? If its price is 1 tick lower than 2nd favs price - is it still a fav?

That's right I simply rank all the runners by price, & the lowest priced runner @ 5mins gets the label of fav.

Do you mean drift trend weakness?

Yes, to be a valid entry the runner had to be trending

[Korattt]

Sorry Euler but don't understand ... "IP"? Not In-Play!

And care to share the degree of %?

Implied probability > (1/odds)*100

Curious though, what advantage does this have over just looking at percent change in odds?

boxing everything into the same common currency (despite odds being the same currency, they vary widely from 1.01 to 1000 odds, so do not show change inside scale very well) means that you more easily determine inter-related moves between runners, even if their odds are fairly widely apart.

i've said it many times on here, but think of it as pistons on a crankshaft, by having the same restraint (the crankshaft and limited length of the piston), you can more easily identify when one runner is moving in one direction or the other. couple that with min/max book% to create outer limits and you have yourself a nice little constrained system...

it might look a bit like this:

book-perc-etc.PNG

Jim,

I’d find that spreadsheet mighty useful, would you be so kind as to share that?