built an orderbook simulator you can read about bids/asks all day, but until you watch a fat market order walk the book and destroy your pnl through slippage, it doesn't click if you're touching perps (or trying to understand Hyperliquid or Lighter) you need to internalize this here's what actually happens when you trade. look at the order book above - every order sits in a queue. bids (green) = buyers waiting. asks (red) = sellers waiting. that yellow line in the middle is that's the "price" - really just the midpoint between best bid/ask spread = best ask - best bid. tighter spread = more liquid market = better fills limit orders add liquidity. i'll place a 0.5 BTC buy at $99k. updated book: see how it just sits there in the book? that's making. you're providing liquidity for others to take maker fees are lower because exchanges want thick books. more liquidity = more volume = more fees market orders are violent. watch this 5 BTC market buy - it doesn't get one price: it eats through $101k, $102k, $103k... your "$100k buy" becomes $101.5k average fill see that volume distribution bar at the bottom? 81% bids vs 19% asks. when this tilts hard one way, market's about to move (unless there's phantom / pvp flow) you can upload your own trade sequences via json to see how different strategies and trades affect the book. if there's interest i'll add perps specific features: funding rates, margin requirements, liquidation modeling. want to show how a single whale liquidation can cascade through an entire book most people lose money on perps because they don't understand the game they're playing. you're not trading against "the market" - you're trading against other participants who can see everything thing you're doing. play with it. upload some trades. watch how the book reacts. then maybe you'll understand why 90% of perp traders get rekt link below

the founder skill tree by default you are bad at almost all things when starting a company, especially if it is your first one. Often you only have only 1 thing you can be considered slightly good at, but it takes more than that to build a startup. So what you need to do is seek a cofounder who has chosen a different part of the skilltree and can complement what you are good at. AKA classic BD + eng cofounder duo. I built a few businesses in high school and made a few grand. I then tried to spend that money paying a dev to build me an app. He took all my money and told me there were too many bugs and never delivered the app. I was pissed because I was broke once again so I decided I needed to have an engineering background so I could at least rely on myself to build an MVP. Downstream of this decision I took a 4 year detour to go study computer science at Harvard. Then when I graduated I realized that Harvard didn't even teach me how to actually engineer systems, it just taught me how to do math proofs and think about computers. So then I basically spent my whole first company trying to learn how to engineer systems despite being CEO. Honestly I got really lucky that Ripple ended up buying it because the only skill path I had leveled up in was engineering. I only had enough skill points allocated to the BD tree to just barely negotiate that acquisition. Then when I started my second company I tried to level up in sales pretty rapidly, and it worked decently well. We pulled $50m into our DeFi protocol within days of launching with 0 public marketing. Then Terra blew up causing a cascade of liquidations nuking this industry to the stone age for 3 years. During this time I should have decided to level up in product, because this really was (and is) the foundation of any successful business and I really was not (still am not) great at it. Instead what I actually spent skill points on was marketing. I learned how to game social algos to distribute content on @OmniFDN and grow awareness of our project. I got pretty good at this -- I drove 25M impressions just from my personal account over the past year, and 16M of those were from just two months where I was really focused on it. None of that matters though if you can't convert those eyeballs into usage of your product, and we didn't have PMF, so no matter how many views I drove it didn't actually really help out with the business fundamentals. Now, after 7 years of building startups it has become painfully obvious that the skill tree branch I needed to max out much earlier in my journey was product. Despite this taking years longer than it should have, I finally feel I am getting at least the foundations right because we are seeing great early signs from the product we recently launched in private. To level up in marketing I was maniacal but pragmatic -- I spent days on end going back to the very beginning of popular accounts (people + companies) on crypto twitter and looked at every single tweet they posted ever and logged the ones that drove higher engagement. After enough hours of this you start to see very clear patterns & tactics that you can replicate for anything you're building. On product the formula that is starting to show results for me is pretty simple, read the following 3 books: 1) 4 Steps To The Epiphany 2) The Lean Startup 3) The Mom Test And simply follow the extremely time intensive process laid out by 4 Steps To The Epiphany. Talk with users constantly. If you're not talking to at least 5 prospective customers / week you're failing. We haven't publicly announced anything about the new product we just launched in private at Omni (soon, but if you are using perp DEXs today my DMs are open if you're curious) but I can feel that the foundation of a great product is being built. We have much to learn as a team, and I have much to improve on as a founder. But at least the domain that I am smashing my head into a wall 12 hours a day feels like the right place to be skill point maxing at this point in my journey. send tweet

perps 101 perps are 90%+ of crypto volume today, they run the market. critical to understand, but most people working in crypto don't actually understand how they work. did a few sessions this week internally breaking them down. here's the tldr - not a guide, just my stream of consciousness notes. start with order books - they're universal, not specific to perps. bids = buy orders, asks = sell orders. difference between best ask and best bid is the "spread" there is no true singular "price" of an asset. it's just the midpoint of the spread. on coingecko or cmc, they take the avg across many exchanges. makers add to the book, takers grab what's there and move price. exchanges want bigger books, fees incentivize this: maker fees are lower, taker fees are higher. post-only orders are a way to guarantee you're making not taking. market orders walk the book. your average fill is always worse than top of book. snap buy then snap sell = guaranteed loss equal to spread plus fees. stops and takes: take profit is just a limit order resting on the book. stops are usually triggers held off-book by the exchange. mature exchanges fire stop losses cleanly. newer ones don't, so set alerts for your stop losses. forwards = agreement to buy or sell an asset for a price on a date in the future. requires actual delivery of the asset. futures = forwards with 1) standardized contracts so there's better liquidity 2) margin is required 3) marked to market daily. futures enable leverage. futures converge to spot at expiration. if the price of the future > price of underlying, arbitrageurs come in. they can do this because price converges at expiration. liquidations happen when margin is close to zero. your position is closed by opening an opposite direction market order. ex: ETH long gets liquidated = exchanges opens a market short on your behalf. this pushes price down further. might trigger more liquidations, which open more market shorts, pushes price further. that's a liquidation cascade. in the other direction, called a short squeeze. people like leverage from futures. but they want positions open for a long period of time, no expiration. enter perps: futures with no expiration. but expiration is the anchor that keeps future prices tied to spot - without it, perp price will diverge from underlying with no tether. wat do? funding rates. if perp > spot, longs pay shorts. if perp < spot, shorts pay longs. arb-ers close the gap. funding payments happen every 8 hrs or 1 hr depending on the exchange. perps are always on. you can hold exposure indefinitely. just watch that your funding payments aren't eating into your returns. funding is often positive (longs pay shorts) bc crypto people are hungry for levered longs. this creates a "cash and carry" arb: long spot, short the same asset perp. rake funding payments. this is what Ethena does. typically collateral is USDC. with the cash and carry arb, there's a possibility of liquidation if prices move against you. some venues let you use crypto as margin. better cash and carry: long ETH spot, use ETH spot as collateral to short a perp. ETH price increases, you need more margin for your short, but your collateral increased in price. chillin. with onchain perps, everyone's liquidation points are known. creates a pvp book. treat it as adversarial by default. perps collapse everything a trader wants into 1 instrument: leverage, simple, 24/7, no borrow friction, marked-to-market. BitMEX introduced the first perp in 2016. trade BTC on leverage. s/o Arthur Hayes. today, they're the center of the market