Hello and welcome to Backtest Tuesday!
Welcome to Backtest Tuesday! This will be an ongoing discourse for everyone to participate in random backtesting, realize market falsehoods and try out things that may or may not work, myself included! We will use the terminal backtesting function (augmented with custom studies) and through the week the thread is open to ANY idea within the parameters of the model.

To recap, last week we added some basic trend and breakout rules to the range-bound strategy and tweaked the exits. The equity curve looks nice, our objective function (profit factor) has been maximized while drawdowns have been lowered. Today we take the next step and move into testing with out of sample data to see how awesome this strat is! drumroll please

Welcome to Backtest Tuesday!
This will be an ongoing discourse for everyone to participate in random backtesting, realize market falsehoods and try out things that may or may not work, myself included! We will use the terminal backtesting function (augmented with custom studies) and through the week the thread is open to ANY idea within the parameters of the model.

To recap, last week we built a range-trading strategy for Russel 2000 Futures using moving retracement levels. The ultimate goal is to try and build an all-in one model that switches between range-bound and trend following.

They key points from the range-bound strategy are:
1. Short at the upper retracement, long at the lower retracement
2. Only when price action is range-bound
3. And only if that range is compressed
4. We stop out if price action exits the established range and losing more than 2.5% at the close
5. We place trailing stops using retracement levels inside the range where short term momentum is weak

Please read last week's thread QT'd here for a full rundown of how we got here.

This week - we will tackle the trend portion! The equity curve of the model with just range-trading is what we expect to see, a general positive performance during sideways periods with pockets of being flat during larger moves (and naturally expanded volatility). The aim now is to fill in these quiet periods with effective positioning to capture trend.

Welcome to Backtest Tuesday! Yes, it's back.

This will be an ongoing discourse for everyone to participate in random backtesting, realize market falsehoods and try out things that may or may not work, myself included! We will use the terminal backtesting function (augmented with custom studies)

Through the week the thread is open to ANY idea within the parameters of the model. To start...

This week we will use Russell 2000 futures. The objective for this week is to find out:
1) Can you build a model that switches between range-bound and trending
2) How do you trade either one
3) Is it even worth it?

For the sake of simplicity we will use the following parameters:
1. A constant capital amount of 100k (no compounding) with 1MM equity.
2. Daily frequency
3. Long/Short
4. 2017-2022 as our in-sample period. At the end we will then review performance using 2023-2024 as our out of sample period. To set up the first part of this I will naively define the asset as being range-bound if it is trading in between two (or more) established levels in a lookback and trending to be anything outside of that.

To set up the rangebound conditions, I am using moving retracement levels built in STDY

The top panel defines the parameters for the lookback period, and 4 variable retracement levels.
The middle panel is to output the moving retracement levels and the bottom is just to output the actual data onto the chart.

The reason why I bring up the neutral rate is because I think the market is actively debating it without realizing it or explicitly saying so.
Much of the consternation around Fed pricing right now is how many cuts is appropriate this year, but that implies that cuts themselves are appropriate. One of the key tenets for this view is that with inflation moderating and (to a lesser extent) activity cooling, the NOMINAL Fed Funds rate, if left unchanged gets more restrictive as inflation falls due to the real-FF increasing (Nominal-FF minus Inflation = Real-FF). Thus, in order to cool the economy without overtightening the Fed has to cut rates.

Economic Syncopation Update: Ummmmmm

Either we have a recession really soon or I think the Fed needs to hike?
Allow me to explain. The previous thread on Economic Syncopation and trusting the data is here for your savage reviewing and scathing criticism.

https://twitter.com/EffMktHype/status/1732241858498298095

Economic Syncopation: Trusting the Data

The key takeaway from my last update on Econopation™ was that if we looked at the rate of change and not levels, we could understand how various eco indicators are evolving and whether sentiment and reality are back in sync...or not. If not already clear, by turning to RoC we want to observe the indicators' direction relative to one another but also whether they are growing or declining (i.e. +/- 0%)
Smoothing out series' like ISM allows us to see the general trend: PMIs are flattening out (heading to 0%)

The reaction in bonds to CPI y'day was everything and nothing at the same time. If you told me inflation was going to come in soft and bonds would rally on the end of hiking cycle, I'd probably think the curve bull steepened.
Yet while everything rallied, the curve did nothing
To be fair, the other curves steepened more, with the belly outperforming the most. But again, it raises the question of why we rallied so hard.

An Update to Economic Syncopation.

I previously conceived this idea of non-synchronization in the economy back in Jan (that long ago??) and how the pandemic disrupted the natural cadence of leading and hard economic data.
I'm happy to report I'm not the first. It was not novel.
For those that missed it, here is the original thread. The basis of the discussion was how the pandemic and subsequent rescue compressed typical reaction times, thus causing everything and everyone to be out of lockstep - syncopation in the economy

https://x.com/EffMktHype/status/1619985221524520965?s=20

Rate vol continues to climb, with long end vol continuing to drive the show. Pain continues to be wreaked in the longer tenors as evidenced by the bear steepening in the curve that has not abated.

So...is it time to start nibbling on duration? Much of the pressure is in 10s but 30s bears (pun intended) watching. The 2022 rate hike campaign that ratcheted front end vol higher from monpolvol has ended. Recall that long end bonds were not spared - but they were passengers.
That the vol complex flipped raises the

This is the ratio of 1 month ATM implied vol on ES vs TY. It recently went as low as 1.61, only seen for a blip in 2014. Other notable observed lows close to this level was 2013 and 2016 (the first clue).
Now if you're expecting bearporn - let me finish first before you do
/ The point of a ratio is that is involves TWO moving parts. Now with that in mind, let's try to deconstruct this ratio and find some sense in it. The obvious questions are

"why is it so low, what does it mean and what does it imply?"

So let's start with the first qn.

"2s' peak always exceeds the peak in FF". Is the pain trade for rates clearly to the upside? Doesn't mean it will happen but it does mean it's possible and the market is still clearly reticent to acknowledge it (which continues to contribute to the divergences cross-asset wise.
/ The idea that 2s always exceed FF in a hiking cycle is something I am guilty of proposing, even back in August when 2s were sub-3.50%

https://twitter.com/EffMktHype/status/1563221345093054464?s=20

/

If ES/SPX closes relatively unchanged on 5/18 you can expect realized 10d vol to drop to ~8% annualized. What's that mean?
Fuck all. As in literally fuck all is going on.
Here let me show you how little fuckery is going on.

F/1 First we set up some variables to help measure fucks. Default it to 2% close to close change for a threshold in fucker and 20 intervals for the fuckery lookback.

F/2

How have things changed over the first quarter of the year? In a way...not much. But the nuance is in both levels and direction.
Sometimes walking in a circle helps you spot what's at the edge of the forest, sometimes running on a treadmill gets nowhere but you're fitter. Case in point: The US 2s10s curve is.......the exact same spot it was Jan 1. You're back at the same spot in the forest.
But what the trees tell you is that there was a persistent pressure for the curve to flatten only for that to be rudely blown up.

March 2023 Gif Recap

Macro traders Regional banks

The BBB Series – Topic 7: Part 2 – Exploring Other Swap Types

Here we will take a brief look at a few other swap types that have been mentioned in this series so far, as well as some swap types you might read about in news/fintwit. Again, we will not go too deep into each of them but more so focus on building the right intuition behind each of them – reasoning trumps rote.

As promised, a thread on my approach to building a rules-based model. There's no perfect way, you won't get a blueprint of what I do, just how I approach a systematic trading model. I hope this provides some value in constructing a strategy, rules-based or not!

Let's go! First, let's define a systematic approach. You're essentially taking defined and measurable inputs and packaging them together to form coherent and consistent rules to establishing positions in various market instruments.
Ever notice how some traders refer to their "system" even

I think that after yday's drama it is worth considering the amount of structural damage that has been done and the likely fact that severe market gyrations aren't over despite stocks higher in Asia this morning.
I'll mainly focus on rates/credit to illustrate. Rate volatility moved insane amounts. TYVIX shifted 4.39 points, the largest single move in a day. FYI, TYVIX is calculated the same way as VIX, but using the US 10 year future (TYA on BBG, /ZN on CME ticker)