Lionel Lightcycle
@0xLightcycle
Followers
9,647
Following
1,574
Media
472
Statuses
4,849
crypto quant | on-chain pvp | rated 9.999916/10 by hot market makers
Joined February 2021
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Iran
• 867734 Tweets
Beyoncé
• 418807 Tweets
Judge
• 378835 Tweets
Joe Rogan
• 309804 Tweets
Dodgers
• 285396 Tweets
Yankees
• 195756 Tweets
#WorldSeries
• 158208 Tweets
#モンスト
• 155697 Tweets
期日前投票
• 132777 Tweets
Freddie Freeman
• 127942 Tweets
開催中のガチャ
• 118741 Tweets
中野家の五つ子のうち1体
• 108318 Tweets
Grand Slam
• 108155 Tweets
ドジャース
• 101675 Tweets
フリーマン
• 89152 Tweets
Lakers
• 84956 Tweets
ワールドシリーズ
• 64230 Tweets
満塁ホームラン
• 52554 Tweets
Ohtani
• 39478 Tweets
Nestor
• 34239 Tweets
Stanton
• 34194 Tweets
Boone
• 31279 Tweets
中野三玖
• 30622 Tweets
中野二乃
• 24467 Tweets
#池ハロ2024
• 22852 Tweets
Mookie
• 18099 Tweets
大谷さん
• 18037 Tweets
Verdugo
• 15406 Tweets
アロンズロッド
• 13587 Tweets
Betts
• 12075 Tweets
Kirk Gibson
• 10842 Tweets
Pinned Tweet
Trade six types of strategies
If you have momentum trades, add in a system for mean-reversion. If you have trend-following trades, add some carry strategies. +5 from one strategy, could be -4 from the other and it'll auto-reduce scale
h/t discussion w/
@SimpelAlpha
Thread👇
19
95
426
If you're not retiring this month, start ramping up your market neutral strategies -- these are carry, pair/spreads, cross-sectional
1/n thread 🧵
10
126
794
Apparently in crypto, I'm suppose to know and use:
Go for MEV
Solidity for EVM
Rust for Solana dev
Python for Stat Arb
C++ for HFT
All at once. Multiple times per day
33
33
475
First step to quant is to trade the markets by hand to build intuition
All we’re ever doing is coding up something that approximately captures the effect
And if the effect isn’t already there, we can’t “math” our way out of it
6
23
234
HFT takes money from urgent traders (orderflow)
Orderflow traders take money from those who demand liquidity (statarb)
Statarb traders take money from long-term allocators
Long-term allocators take money from macro
Macro traders take money from the future (unborn children)
10
16
176
What is cross-sectional momentum and how would you build a crypto portfolio with this?
🧵
3
32
159
If you're just starting off on-chain, automate a DEX swap with ethers.js first
Don't jump directly to contract deploys to testnet/mainnet with Solidity
General learning plan would be:
1/x
3
20
158
@functi0nZer0
People should just stop naming their companies Bankrupt if they want to stay in business
2
1
145
First it was crypto-to-crypto beta
Then it's crypto-to-equities beta
Now we're talking about Japan and crypto-to-fx beta
Starting to sound like we're just trading tradfi for 2023
3
6
128
In reality, you better have great relationship w/ the exchange (read: you give them lots of $)
Cause they'll tag you for your quote-to-fill ratio
1
0
125
Time is the easiest way to group bars, and is noisy if you're not specifically trading a time-based strategy (e.g. buy BTC MWF, or snipe Asia bids)
Consider an alternative grouping: Dollar Bars -- e.g. create bars every $1B
Daily vs. Dollar bar comparison in image below:
1/n
6
13
125
So interesting getting RT's of old SBF tweets and seeing them in a new light
It's like sitting through a re-watch of The Usual Suspects or Memento
7
7
126
If you hate someone, tell them Kalman filters are the secret
If you don’t, try these first
6
9
114
Almost back from break, been spending most of it reading, old finance books & some fiction
I use a trick to consume literature the way my PhD friend taught me:
Build an index of the content in your mental ram, and then use multi-pass-reading to tie the concepts together
1/
3
11
120
c++ HFT codebase here:
A roadmap on how to improve it from
@TerribleQuant
:
1. Modify to use asymmetric spreads
2. Introduce a fast ML model
3. Implement microstructure indicators
getting out the way of adverse conditions by making your spreads asymmetrically wide (traders with alpha against you).
1
1
13
6
14
115
I trade with silence, and I feel like a weirdo because everybody is blasting techno music in their trade streams
17
5
100
@lostbutlucky
This could be either a pitch for a YC-backed dating startup, or a reality TV show
3
2
94
Market orders are optimal when inventory and depth imbalance are at extremes, to add more flavor to
@IDrawCharts
point
1/3
remember kids:
if you're bidding a level and betting on a pullback, if the market is strong you're unlikely to get filled and will probably get left behind when the market rises
but if the market is weak and you misread it, you're guaranteed to get filled when the market drops
12
17
136
1
19
95
It's a deadman's switch
I'd send cryptic things to Twitter too, to trigger actions
7
3
89
Before reaching for that clustering algo, build an intuition about the market
Ways to do this are:
1. Animations
2. Cross-sectional heatmaps
3. Sector studies
See if you can spot the sectors and high/low market cap patterns in the animation below
HQ:
4
6
92
An equal-weighted ETH/BTC outperformed a beta-hedged ETH/BTC construction during recent market move
This is due to the difference between beta, up-beta, and down-beta
When would we want to use an equal-weight vs. beta-hedged?
First, will do a few non-mathy definitions:
1/
3
18
88
@worstcontrarian
They’re actively inventing a demographic that the readership will not identify with
This is how WSJ creates us vs them narrative:
1. pick a young guy (their readership skews older)
2. Say the guy lost <$1k USD (like slot machine amount)
3. Make the guy a crypto fanatic
4
3
83
Basically, this:
1
3
85
Looking into new candlestick plotting libraries in Javascript
Apex
HighCharts
Or the new kid on the block
@ViperCharts
8
9
85
Taking a stab at this.
@M1tchRosenthal
kicks-off interesting points in the thread
The "entries vs. factors" discussion can be thought of as:
1. Entries/Exits are complete as a trading style
2. Entries/Exits (Before) evolves to Positions/Factors (After)
So, which one?
1/
Sometimes, I think we make too broad statements about things in quant land.
For example, many institutional practitioners argue that thinking about entries & exits is bad.
If you are trading based on factors, makes sense. But is trading factors the only way to quant trade?
4
5
27
1
20
82
This PDF can be a framework for R&D and implementation of a profitable trading strategy
Fill in the blanks where it’s redacted, and it walks you through identification, research, environmental (heuristics), triggers, improvements to execution, and so on
3
1
81
Follow the smaller accounts
Their risk-adjusted follower count is off the charts
Not a single down day in months
6
1
78
People would improve on tick/DOM knowledge much faster if counterparties could make fun of you on live chat
2
2
77
I dreamed about an edge, and it was vivid enough to show the data required, the horizon, environment, the market effect
Woke up, heart pounding, ran to my desk to code it up
When I was done, the data was crystal clear in its message:
“This is garbage”
Have a mid 20’s engineer at my work. She changed her mind on a technical decision after a good nights sleep and was worried she was flip flopping and I had to coach her through the “technical decisions made in a dream are insanely Lindy trust the spirit world”
42
172
4K
6
6
79
This kind of signature shows up all the time
The majors get liq'd in a short squeeze (BTC open interest down), and then alts just market buy according to their beta
3
1
75
I found this difficult with high volume, an easier way to get into the zone is:
1. Don’t start with staring at BTC/majors
2. Train your brain to identify counterparties, aggro, wide quoter, liquidity chaser
3. Categorize environ against
#2
with tools or brain
4. Then move faster
4
3
75
Here's how I think of data collection for research purposes:
1. Flat-file Data lake (e.g. S3/Filesystem)
2. NoSQL Data lake (e.g. InfluxDB/MongoDB)
3. Data warehouse (e.g. TimescaleDB/Clickhouse)
Depending on the data...
1/
6
11
73
If you guys want a visual on why Dec 1st is important, take a look at the bid liquidity
You could see this days in advance that it's widening (thinner books, so less effort to blast through)
Also, go check out
@Mtrl_Scientist
thread for great explanation
5
10
64
The pure news traders basically already out within 1hr
3
5
64
You'll that a quant alpha like autocorrelation or an effect fading, I find that applying it discretionary can still eke out performance
e.g. with a basic automated regime filter, it'll start failing, but there's something in experience, intuition, that can make it work
1/4
1
10
71
The open interest of the CT users who shoot first and ask questions later is about $6MM
This bit of info is going into my altcoin models for the future
2
0
69
Hsaka's preference for executing the higher conviction side (i.e. no pair trade) is optimal when the edge is picking the strength of a ticker
You don't need to pay a "hedge cost" to add a bet you didn't want in the first place
2
6
70
It's that kind of market -- the majors are moving and the median alts are relatively flat
2
4
67
If you see BTC and the market mean move like this, just start doing TA on the Open Interest
6
5
67
Quant Trading and Market Making with
@AlamedaTrabucco
Interesting point on decision making at the firm: weigh own decision 100% or 0% depending who is more knowledgeable; example w/ how to play the OMG/BOBA fork
h/t
@trader_hades
1/3
2
9
69
Been experimenting with the spike and decay function to get out of convergence trades
I think any application that desires a trailing stop can benefit from it
I see parallels between the half-life and the distance of a trailing stop
6
2
68
People spend so much time talking about data analysis and not enough about market effect
You can improve a little in your prediction with data tricks
Or you can just model the effect correctly in the first place. This requires experience in trading
A quick example with OI
1/
1
9
69
Use State Machines for Trading
The way to mess up the design is to:
1. Make events emit events
2. Deeply nest the state machines
While not a pure FSM, the cleanest way to use state machines is to keep money making logic at the top level — called “inverted”
Few more resources:
Thoughts on writing trading strategy monetization:
The lower latency you go, the more tedious the state management gets. If you can do your trading with a CSV upload of market-on-close orders to your broker, do it! If it needs a broker algo every hour, do that.
3
17
178
3
8
67
Fading the edge of range works every time until it doesn't
Shorts who accumulated over past few days closed out on that move
7
1
63
Quant: spend enormous effort to neutralize beta, tight volatility controls, trading hundreds of tickers, manage txn costs
Trader who makes money: buy when the coin goes up
I love that most markets are traded largely on fundamentals
I trade on what is going up and down (plus personal biases) and I beat all of them lmao, just give me a trade feed and an orderbook
couldn't care less what the ticker is
5
1
88
8
1
68
People sending marked up charts and asking for input with no other context
This is backwards
Market review goes in this order:
1. Environment (Macro, sentiment, your favorite twitter account shilling)
2. Scanner / Overview (rel value, Z, beta)
3. Execution (chart)
1/
1
7
65
Agree with more focus on analysis tooling vs trading automation
Have had experiences in wasting potential of profitable strategies due to automation time investment
One trick is to create signals that span markets instead of a specific extraction, gives a few benefits like:
1/
4
5
65
@bitbitcrypto
Everybody wants ride or die friends, but you never know until the 2am call goes out asking for shovels
1
3
62
Honestly relieved now to be zeroed out by the hacker and can just stop looking at it
7
4
63
You can’t just pick two random tickers and call it a pair trade, there needs to be some sort of reason otherwise you’ll bail at the first sign of volatility
8
6
58
Gave advice in DMs about quant and they replied with a 3 paragraph strategy for League of Legends to show their appreciation
I don't play League and have never talked about it
Much appreciated all the same, will reference this for my pro gaming career in the future
3
0
57
Finally tried out Foundry after you guys recommended it
The first thing I did was build the Uniswap core and periphery contracts. No issues there.
The next thing I did was write a contract that interacts with Uniswap -- and ran into a bunch of compiler version conflicts
1/x
4
4
62
A friend asked me for trading help, which I refused on principle
I eventually gave in after I learned they were caught up in the UST depeg, so I taught them the OI shed / liquidation trade setup
My thinking: hopefully they wouldn’t lose that much money and get screen time
1/2
3
2
59
@macrocephalopod
Zillow business plan leak:
Step 1. Tell everybody when and how much you're going to buy
Step 2. ?
Step 3. Profit
2
3
59
ex hft: used an API once
quant: plotted BTC against ES
cto: made a crypto-themed todo app
vc: paid for netlify and got discord mod
20 yr olds on linkedin be like “ex hft, quant, cto @ startup, vc”
bro how you cant even drink
29
18
343
2
2
58
No matter how much I trade, I can never beat the performance of some wallet that I forgot about
9
2
55
We're ramping up, if you know a fit please DM Lily
Does not have to be crypto specific, since trading is a mindset and not an asset class
Look forward to talking to you
4
2
58
Telling myself I'm too stoic to be affected
Inside
I'm crumbling
4
0
55
In the span of a few hours, let's liquidate the longs, then liquidate the shorts
5
1
49
If you’re not doing high frequency and think you have A, you have the inverse of C
imo there 2 arguments against the very valid concern of edge erosion:
1- Sharing is caring, you tend to get back more that what you put in.
There are luminaries sharing fantastic content and it accrues in unmeasured intangibles - goodwill, opps, and an increased luck surface.
13
33
296
5
0
57
Eventually, a whole set of newcomers enter the fray, and things that used to work, but stopped working, but might work again are:
- low unit bias
- round numbers
- round lots
- meme numbers
- sector leader chasing
6
3
53
Is "see you on the orderbooks" friendly or aggressive?
cause I've been using it a lot in a friendly way and am just realizing it may not come across that way
13
0
53
These cringe memories pop in my head and keep me up at night
Like of my emo-quant phase with MATLAB and wanting to use the Kalman filter for everything
3
0
54
Only someone who stole it would do something like this
Someone just dumped 400+ BTC over 2 hours in 10-50 BTC clips on the XBTUSDT pair on Bitmex eating 30%+ slippage. They must've lost $4m+ at least.
157
224
2K
4
1
53
If you only have one server, put your server in AWS Tokyo
Close to FTX, close to Binance, kinda close to Bybit
Plus, you can whitelist your FTX server IP to skip Cloudflare if you're in Tokyo
3
3
51
This area open interest gets spooked easily, each way
New OI (breakouts) is what pins levels
9
5
52
Db? Postgres
Queue? Postgres
Blob storage? Believe it or not, also Postgres
We have the best quants in the world. Because of Postgres
you can just do it all in postgres, nobody can stop you
29
9
187
6
4
54
Flowchart on portfolio weighting based on your views of the market:
1. Start with equal-weighting if you're not sure
2. Volatility-weighting is surprisingly divisive
3. Correlation-weighting if you are betting on correlations specifically
You can mix types of weighting too
3
5
51
@worstcontrarian
And now we’re attacking this invented persona instead of the media
This is literally psyops
1
0
50
@0xrush
Trend Following = Time-series momentum
Mean Reversion = Reversal momentum
Carry = Liquidity momentum
Pair/Spread = Convergence momentum
Indexing = Beta momentum
2
4
51
I'm thankful for services that serve special data on interval APIs only, so I can enjoy the edge before it's gone
You may want to reach for a headless browser to do this, but you only need a few tips to pull it with standard library tools:
1/x
4
5
50
Is there is anything better than
Looking to migrate my whole system to this but want to hear horror stories first
3
5
54
Podcast w/
@macrocephalopod
Points around accounting for "knowns" to surface hidden info stood out to me:
- example on hedging away correlation and increasing lev
- finding proxies to snapshot events to tune exposures
- surfacing signal by netting away known fundamentals
1/
Managing a Mid-Frequency Crypto Prop Desk with
@macrocephalopod
🪦 Is naive style premia dead?
⌛️ Errors in backtesting
🚥 What is "mid-frequency"?
🏗️ Building a desk from scratch
🌊 Momentum as a risk factor
💥 Impact, fill ratios, and borrow rates
9
23
208
2
2
52
You trade with method X? Here's why X sucks
Method Y is the best, subscribe to my newsletter to learn more
2
4
53
Emphasize *layers* of security rather than yet another service to save passwords
Layers means:
- Top Secret (Offline): Paperkey or encrypted USB w/ Tails
- Secure (Local): Keepass or Passwordstore
- Convenient (Cloud): Self-hosted Bitwarden
Links below
1/
2
11
51
Build your own API server to make your life easier for research and production
Live trading needs backfills that span further than exchange limits -- solve this once and as use the same endpoint for research
Here's what my time-bar wrapper looks like:
1/n
4
3
52
This guy has a hotkey setup for trading, and here I am entering numbers, memorizing CLI args, and clicking checkboxes like a neanderthal
Shout out to
@ichimikichiki
and his epic work on
@ichibot
!
#Ichibot
is such a powerful tool your imagination is literally the limit with the things you can do with it!
NGL combining this legendary order execution tool with
@elgato
's
#streamdeck
brings it to a whole new lvl.
17
12
77
3
3
51
I'm seeing two ways people capture momentum in single names:
1. trigger on the dip
2. trigger on the breakout
Of the two, the breakout gives more information about the market
3
2
48
A few ways to play these launches
First phase:
1. Listing giga pump
2. Fade the blow-off
3. Spreads / Arbs between exchanges and Spot/Perp
4. Short every perp listing
Second-phase:
1. short squeezes
2. liquidity triggers to test ranges
Then: bleed or forced narratives
3
8
47
There are two kind of pair trades:
1. Narrative
2. Structural
Seeing people build narrative trades and then treat them as structural
Structural pairs trades thesis more clear, like: gov vs. rewards token (e.g. Gaming), high vs low inflation w/ same tech (e.g. Solana)
1/
4
5
49
You know how Serum is basically Wintermute et al. picking off Alameda on Solana?
Apparently, FTX orderbooks are the same thing (at 10B+ scale), where everybody picks off Alameda on FTX
1/ Very rough and speculative sketch of what I increasingly think happened at FTX as more info comes out…
The central question is where did the money go? Yes malfeasance and fraud is necessary, but at one point in the cycle cash actually has to go out the door
335
744
4K
9
2
47
A non negligible amount of my ideas have come from trying to extract something from the market and getting beaten.
Then front running that event.
Then getting front run.
The front running that.
Until nobody makes money anymore
2
4
50
You can either complete both legs of the trade and risk prison like Abraham
OR just short the perp and tweet about the opportunity, let someone else complete it
The imprisonment risk is only for the illiquid leg
So people are manipulating Comp because the thin liquidity makes it super easy (just ~$150k spot buy pumps it like 1% lol).
But do they realize I can borrow $5M comp onchain right now? Shorting perps then dumping all of this would be an easy 25% nuke. Hypothetically, of course.
40
17
327
2
5
49
Visualization of Top & Bottom pairs by Z-Score
Tight groupings on the negative side is due to up/down beta differences
Even in this holiday period, can see relative value holding up
... there are people still trading?
5
4
49
When I followed up, they’re sending me screenshots of their TA ideas on BTC-MOVE
because the setup I showed them was “too boring”
2/2
4
0
48
Hackers waited for Tornado announcement to exploit Curve frontend, prime time to capture the escape from USDC swaps
3
2
45
CEL market is fun to watch
The lending spikes several times a day, and then within a bit, participants get squeezed, OI drops, and perps run up
Almost every time
2
0
47
Paleologo's book is my favorite, he recommends Isichenko's
5
0
46
Random snapshot of alts
Notice how alt moves are liquidations and the majors moves are not (CVD up, low liqs)
4
2
46
Worst trade in my career was expecting an arb to last
Didn’t size up enough, because I spent time writing code instead of continuing to slam it by hand with an alert
And opportunity gone soon after
The worst trade in my career was something that looked like an arb but wasn’t.
These types of trades are toxic, and they can blow up your account because of how easy it is to size them up.
8
3
108
6
0
47
Time to start market making. Been avoiding it b/c of tech, but I'm seeing spreads that appear more lenient
What should be taken into consideration when deciding hedged vs. unhedged mm methods?
Hedged mm is easier to code, so I'm tempted to start with that
5
3
47
Strategic momentum:
- Turn off the top one
- Add size to the bottom one
Strategic reversion:
- Double down on the top one
- Turn off the bottom one
Make Strategic Momentum a thing 2022™ h/t
@hhawnk15
3
4
44
Everybody knows "crypto market maker" doesn't mean MM in the traditional sense, it means a whale is gonna pump your coin and wash trade it so it has a lot of volume (to draw people in)
Didn't say it out loud back then since "any activity is good activity"
Still the case today?
YGG playing out pretty ugly
perps came down very hard last 2 hours, currently trading ~30% below spot
3
1
20
5
5
43
Just switched on the long-only (flat positions allowed) basket for a few alts
Did I backtest it? Yes.
Did I ignore the backtest and run it anyway? Also yes.
6
0
44
@macrocephalopod
The other competitive strategy when time-limited seems to be "take high variance trades and pray"
1
1
45
No way
@taylorswift13
for President! I just saw her at her concert in Singapore and realized that she can bring together Americans and people in most countries much better than either of the candidates, and that bringing people together is the most important thing.
Watching this
970
567
6K
3
2
44