8-year veteran's 3-exchange playbook

1. The full picture (and what's not on any exchange)

Before I get to the exchange split, the most important number isn't 60/25/15. It's the split between exchange-held capital and cold storage. Roughly half of my total crypto net worth sits on Ledger hardware wallets that have never touched an exchange. The 60/25/15 is the active trading bag — the part I actually want available for trades, copy trading, new-listing snipes, perpetuals.

The cold storage exists because Mt. Gox happened in 2014. Bitfinex was hacked for ~120,000 BTC in 2016. QuadrigaCX lost $190M in user funds when the founder "died" in 2019. FTX collapsed in November 2022 with billions in customer money. The pattern repeats every 2-3 years. Anyone whose entire bag sits on an exchange is making a bet that this exchange, in this cycle, doesn't get added to that list.

Hardware cold storage is the seatbelt. Active exchange capital is the part you drive with. They serve different purposes and you need both.

2. Binance gets 60% — here's exactly why

The biggest slice for the most important reason: depth. When I size up a BTC perpetual position to anything material, Binance is the only exchange where the order book absorbs it without significant slippage. OKX is half as deep. Gate is one-tenth as deep on majors. That single fact dictates where my serious capital goes.

What actually runs on Binance:

• BTC and ETH perpetual positions. These are the bulk of my active trading P&L. 3-5x leverage default, occasional 10x on high-conviction setups. Never higher.
• Spot BTC and ETH for swing exposure. Sized to be additive to perps when I want directional exposure without paying funding rate (see my funding rate article for the math).
• BNB held for fee discount. About 25 BNB worth — enough to qualify for VIP discount tier on perpetuals. Not a position view, just a fee mechanism.
• P2P fiat ramps. When I add fresh USD to crypto, it goes through Binance's P2P. Best counterparty depth, most options.
• Launchpool and Megadrop participation. Free yield on BNB through Binance's promo programs. Occasionally meaningful, never large in absolute terms.

I don't run copy trading on Binance. Their Lead Trading product exists but the discovery UI is bare-bones compared to OKX. I also don't run altcoin perps on Binance — when I want altcoin perp exposure, OKX is usually similar depth at lower retail-tier fees.

3. OKX gets 25% — the all-rounder slot

OKX is where I run anything that isn't pure BTC/ETH directional trading. What's there:

• Copy trading allocation. About 10% of my OKX balance follows 3-5 vetted leaders (see my month-long copy trading test). Treat this as satellite exposure, not core strategy.
• Options. OKX has the most complete options product of the three. I run modest size on BTC and ETH options — usually selling cash-secured puts and covered calls for premium income. Treated like a yield strategy, not directional bets.
• Altcoin perpetuals. When I want SOL or AVAX or similar perp exposure, OKX is usually slightly tighter on retail-tier fees than Binance, and depth is "good enough" for my sizing.
• OKB held for fee discount. About 1,000 OKB. Same logic as BNB on Binance — pure fee mechanism, not a position view.
• L2 chain bridging. When I need to move USDT to or from an L2 (Arbitrum, Optimism, Base) for DeFi exposure, OKX has the fastest L2 support of the three.

4. Gate gets 15% — the venture allocation

Gate is the smallest slice intentionally. Everything on Gate is "money I'm willing to lose." The strategy here:

• New-listing snipes. Following the playbook in my Gate day-1 piece. Each new-listing position is 1-2% of my Gate balance, which means 0.15-0.30% of total net worth per snipe. Multiple positions running concurrently, capped at 10% of Gate balance total.
• Long-tail altcoin spot. A handful of positions in altcoins that exist on Gate but not yet on Binance. Held with the expectation that they'll either run before reaching Binance, or go to zero.
• GT held for fee discount. Small allocation, mostly because it doesn't matter much — Gate's token-based discount mechanism is the weakest of the three.

I don't run perpetuals on Gate. Their depth is too thin for my sizing. I don't do copy trading on Gate either — the product exists but the leader quality and screening tools lag behind OKX significantly.

5. How allocation gets rebalanced

The 60/25/15 ratios drift constantly as positions gain or lose. Rebalancing is mechanical:

• Quarterly review. First weekend of January, April, July, October. Pull balances from all three exchanges and the cold storage. Compute current ratios.
• Rebalance threshold: 5 percentage points. If Binance has drifted to 55% or 65% of the active trading bag, I rebalance back toward 60%. Under 5 points off, I leave it alone.
• Mechanism: normally I move USDT between exchanges (TRC20, see transfer pitfalls piece). Sometimes I close a position on the over-weight exchange and open a similar one on the under-weight exchange.

This is essentially the same logic as the "rebalance your 60/40 stock-bond portfolio" advice from traditional finance, applied to exchange exposure. Mechanical rebalancing forces you to take profit from winners and add to losers — which is uncomfortable but works.

6. Strategy-level allocation (the layer underneath)

Within each exchange, capital splits into strategy buckets. Approximate numbers:

• Binance: 70% on BTC/ETH spot + perps directional, 20% on BNB for fee discount, 10% in idle USDT for opportunistic deployment
• OKX: 40% on options yield strategies, 40% on copy trading + altcoin perps, 10% on OKB, 10% idle
• Gate: 50% on new-listing slot rotation, 30% on long-tail altcoin spot, 10% on GT, 10% idle

The "idle USDT" line in each is the dry powder for when something interesting shows up. Without dry powder, you can't act on opportunities — and the most expensive trades are the ones you missed because you were fully deployed.

7. What this allocation gets wrong

I want to be honest about the parts of my allocation that I'd change if I were starting over:

Over-weight Binance. 60% on a single exchange is too much concentration for a true risk-isolation play. I justify it with the depth argument, which is real — but a 50/30/20 split would be more defensible from a pure risk standpoint, at the cost of accepting more slippage on large perp trades. If I had started today, I might run 50/30/20.

Under-weight DeFi. Nothing in the allocation above sits in pure DeFi (Aave, Compound, Uniswap LP). I have a small DeFi book on the side but it's not large enough to matter to total returns. With Coinbase Earn-style products and institutional DeFi maturing, the rationale for keeping DeFi at zero is weakening.

No regional diversification. All three exchanges are global crypto venues. If a coordinated regulatory action took down "all global non-US exchanges" (an unlikely but non-zero risk), I'd have a problem. A future allocation might include a small slice at a US-regulated exchange (Coinbase, Kraken) just for the regulatory hedge.

8. What this isn't

I'll close with what this allocation does NOT include:

• Not a recommendation. It's my allocation, built up over 8 years of specific strategy decisions. Copying it directly without understanding the strategies behind each bucket is worse than not having an allocation at all.
• Not static. The split has shifted multiple times over the years. In 2021 I had more on Binance perpetuals (closer to 70%) because the bull market favored aggressive directional exposure. In 2022-2023 I dropped Binance to 55% and increased Gate to 20% as new-listing alpha became more interesting in the bear. The current 60/25/15 reflects the 2024-2026 market structure.
• Not a starter kit. If you're new to crypto with less than $5,000-10,000 total, having accounts at three exchanges is unnecessary overhead. Pick one (Binance or OKX, depending on your region), buy spot BTC, learn for a year. The three-exchange split only starts making sense when your strategy mix demands it.

Real allocation comes from real strategy. Don't try to reverse-engineer mine. Build your own, starting from "what am I trying to do" and "what risks can I actually absorb." The exchanges are tools. The allocation reflects which tools you use for which job. There's no shortcut.

9. The 8-year arc — how I actually got here

I started in 2017 with $3,000 split across Coinbase and a now-defunct exchange called Cryptopia. The Cryptopia portion got hacked away in January 2019; I recovered partial value through the New Zealand bankruptcy process two years later, after paying lawyer fees that consumed most of the recovery. That was the most expensive education I received in my first three years, and it set the operational discipline that shaped everything that came after.

2018-2019: deep bear market. I was down about 65% by January 2019. Most of my friends who'd started with me had quit. I survived by ratcheting down position sizing aggressively — what had been $300 trade sizes became $30 trade sizes. The smaller positions kept me engaged without further bleeding the capital. By the end of 2019 I was back to roughly breakeven, mostly through patience rather than skill.

2020 COVID crash and recovery: the March 2020 -50% Bitcoin drawdown taught me everything I now believe about position sizing. I had a perp position open into the crash because I'd convinced myself the COVID news was already priced in. It wasn't. I got liquidated. The lesson — "your conviction about a thesis doesn't reduce the volatility, it only changes how you'll feel during it" — is one I repeat to myself every time I'm about to size up. The recovery from March to December 2020 was the first time I made meaningful money in crypto, and it was almost entirely from holding spot through volatility I'd previously have panicked through.

2021 bull: I scaled the perp book aggressively as Bitcoin ran from $30K to $69K. Made and gave back several times my starting capital in the process. The lesson here was about exit discipline: I sold a portion at $42K, held the rest, watched it run to $69K, then watched it round-trip back to $34K. The unrealized peak gain I gave back was substantial. Now my exit discipline is mechanical — predetermined percentages out at predetermined price levels, regardless of how I feel about the move at the time.

2022 collapse: Three months of progressively worse news (Terra/Luna in May, Celsius/3AC in June, FTX in November) tested every aspect of operational discipline. The position I was proudest of that year was a 20% cash position I'd been building since the 2021 high; the position I was most embarrassed by was a 7% allocation to FTX FTT tokens that I lost almost entirely. The net result was a 38% portfolio drawdown — painful but recoverable. The 2018-style 65% drawdown wouldn't have been recoverable on a larger base; the 2022 38% was. That's what position sizing buys you.

2023 recovery and 2024 ETF rally: the biggest learning was about emotional regulation during periods of recovery from major drawdowns. Reaching new highs in 2024 after the 2022 trauma was psychologically harder than expected. I trimmed positions too early on several occasions because the "this can't keep going" feeling was strong. The 2024 ETF spot Bitcoin approvals (BlackRock's IBIT crossed $10B AUM faster than any ETF in history) reshaped the market in ways I'm still adjusting to. The 60/25/15 current allocation reflects that adjustment — more spot, less perp, more long-term oriented.

10. The strategy stack underneath the allocation

Allocation percentages are surface metrics. The real story is the layer underneath: which strategies I run at each exchange and why.

Binance strategies. The 60% on Binance breaks down roughly as: 35% spot BTC and ETH held for trend-following with weekly rebalancing, 15% perpetual swing trades typically 3-14 days in duration, 5% funding-rate cash-and-carry when funding spikes above 0.05% per 8 hours, 5% spot accumulation in 8-12 strategically chosen altcoins held for 6-month positions.

The cash-and-carry deserves explanation. When the BTC perp funding rate spikes (typically during sentiment extremes), I short the perp and buy equal-notional spot. The combination is delta-neutral and earns the funding rate as long as I hold. In the most extreme 2024 funding spikes (early March, late November), this strategy paid 15-25% annualized for several days. Across a full year it might contribute 2-3% to total returns. Small but reliable.

OKX strategies. The 25% on OKX is mostly the copy-trading portion (15%) and the options portion (10%). The copy-trading thesis is that I can't be awake and alert across all crypto trading hours, but proven traders with multi-year track records can fill the gaps. I run a curated portfolio of 4 OKX copy-trade slots, sized small (no single follow above 4% of total capital). The combination produces a roughly market-neutral additional alpha source that's loosely correlated with my own trading.

The options book on OKX is conservative: mostly long-dated covered calls on the BTC and ETH I hold spot. Selling a 30-day covered call at 15% out-of-the-money typically produces 1-2% premium income that month. The cost is capping the upside if the spot rallies through the strike, which has cost me roughly 2-3 trades per year. Net positive over an annual cycle but emotionally trying in months when I get called away on a position I wanted to keep.

Gate strategies. The 15% on Gate is entirely about new-listing speculation. I have a watchlist of 30-50 tokens listed within the previous 14 days, and I take small positions (typically 0.3-1% of capital each) in the ones that pass a basic quality screen. The screen filters out clear scams, anonymous teams, and tokenomics with cliff unlocks in the first 90 days. Out of 30-50 candidates per quarter, about 8-12 pass the screen and get capital; of those, roughly half are profitable. The half that aren't get cut at the predetermined stop-loss (typically -30% on the entry). The half that are profitable get held until trailing stop-loss triggers (typically 25% below the local high).

This is the most operationally intensive of the three strategy clusters. I spend 4-6 hours per week on Gate alone — reading whitepapers, checking team backgrounds, monitoring the position book. The payoff has been worth it because the new-listing premium in 2023-2025 has been unusually consistent. Whether this persists into 2026 is genuinely uncertain; the strategy I'm using today was substantially less productive in 2018-2020.

11. The pieces I'd add if I were starting from scratch today

If I were a 2025 newcomer building the equivalent of this 60/25/15 allocation from a clean slate, three things would be different.

First, I'd carve out 10-15% for a US-regulated venue (Coinbase Pro or Kraken) explicitly for the regulatory hedge and the IRA-compatible ETF exposure. The 2018-vintage version of me dismissed this because the fees were too high; the 2025 version recognizes that the regulatory clarity premium is worth paying when you're building toward generational holding rather than short-term trading. The ETF exposure inside a Roth IRA in particular is a compounding structure that's hard to replicate elsewhere — the 2024 ETF approvals genuinely changed what's possible for US-resident long-term holders.

Second, I'd reserve a small allocation (3-5%) for self-custody DeFi positions on Layer-2 Ethereum. Aave, Compound, and Uniswap LP positions have matured enough that the smart contract risk is bounded for established protocols. The yield profile (4-7% APY on stablecoin lending in 2024-2025) is competitive with CD rates while keeping the capital crypto-native. The complexity cost is real but the diversification benefit (no exchange custody) is meaningful for someone building from scratch.

Third, I'd treat hardware-wallet cold storage as a higher-priority allocation slot from day one. My 8-year arc had a 3-year gap where I was running everything on exchanges, accepting the custody risk because hardware wallets felt like overkill. The retrospective lesson is that they're not overkill — they're cheap insurance. A $150 Trezor Safe 3 or Ledger Nano S Plus for the long-term holding portion of any allocation above $10K is a no-regrets purchase. Starting fresh, I'd buy the hardware wallet before I funded any exchange account.

12. Things I'd cut from the allocation if forced

The flip side: if I had to compress the operation to a smaller team-of-one footprint, two things would go first.

The Gate new-listing book is the highest-effort lowest-stability slot. If life got busy enough that I couldn't dedicate the 4-6 weekly hours of due diligence, I'd zero it out and either roll the 15% into Binance spot accumulation or into a passive index ETF. The new-listing edge depends on actually doing the work; treating it as a casual side-bet would produce worse than zero returns net of fees.

The OKX copy-trading slot would be next to go. The thesis there is real but the implementation requires monthly recalibration — checking each follow's drawdown, replacing the ones who've gone off-strategy, sizing up the ones with strong calibration. Without that monthly maintenance, copy-trading defaults to underperforming buy-and-hold because losing traders persist in your portfolio longer than they should. If I couldn't do the monthly work, I'd zero it out.

What I'd keep at any complexity level: the Binance spot BTC/ETH core (the 35% of total that's the long-term thesis), the hardware-wallet cold storage portion (now growing toward 50% of net worth), and the cash buffer that lets me size into volatility opportunistically. Everything else is optional optimization that requires ongoing attention; these three are the foundation that compounds without constant intervention.

One closing thought on time horizon. The 8 years it took to build to this allocation included two genuinely bad years (2018 and 2022), two genuinely great years (2017 second half and 2020 second half), and four years of slow grinding. Returns were not evenly distributed. The years that mattered most for the long-term outcome were the ones I survived, not the ones I won. If there's a single piece of advice that distinguishes the people still here in year 8 from the people who quit in years 2 or 5, it's that surviving the bad years matters more than maximizing the good years. The allocation above is structured around that priority — every slot is sized to be tolerable through a 60-70% drawdown without forcing me out of the strategy.

If you take only one thing from this article, take that. Exchange allocation is downstream of strategy. Strategy is downstream of risk tolerance. Risk tolerance is downstream of honest self-assessment about how much volatility you can actually live through without making the kind of emotional decision that destroys long-term compounding. Get the self-assessment right first; the rest is implementation detail. The 60/25/15 is what implementation looks like for one specific person with one specific trading style developed across two market cycles. Yours will look different. The principles, though, are largely shared: survive first, optimize second, and respect the long tail of operational risk that no leaderboard ever captures.