Weekly Drill Report: 589 Tokens, 32 Survived, and the Selection Rate Just Tripled

The selection rate just tripled. That's not a typo.

Last week I reported a 1.89% selection rate, the lowest the algo had ever recorded. This week it jumped to 5.43%. Same filters, same criteria, same everything. The market changed, not the algo.

589 tokens completed 24 hours of tracking. 32 passed all filters. 513 got hard rejected before they ever had a chance. And the best pick of the week, a coin called $durr, hit 5.78x in two hours.

Here's what that actually means.

The Headline Numbers

589 tokens tracked this week, down 14% from last week's 688. Volume is still shrinking. Fewer tokens launching, fewer people deploying.

But the quality? Way up.

32 tokens passed all of Drill's security and market health filters, giving us a 5.43% selection rate. Last week that number was 1.89%, 13 out of 688. The week before, similar territory. This is the highest selection rate since mid-March.

The median outcome for selected tokens landed at 1.74x. That means half the picks did better, half did worse. It's not a 10x factory. It's an honest number that tells you: the algo surfaces tokens with real potential, but most of them are modest wins, not moonshots.

87.1% of everything tracked got hard rejected. Failed security checks, failed supply analysis, flagged and removed before any user ever saw them. That filter wall is what makes the 5.43% meaningful.

Why the Rate Tripled

Fewer tokens launched. That's the simplest explanation.

When you go from 688 to 589 tokens while the number of quality launches stays roughly the same, the ratio improves mechanically. The trash thinned out more than the signal did.

There's a macro layer here too. Bitcoin is sitting around $67K, down 47% from its October high. The Fear & Greed Index has been parked at 8 (extreme fear) for over a month. Trump's tariffs slammed risk assets in February and the bleeding hasn't stopped. SOL is hovering near $80, and Solana's 30-day DEX volume dropped to $55.5 billion, the lowest since September 2024.

When the market gets scared, casual deployers disappear first. The scammers and serial launchers slow down. What's left is a smaller pool with a higher concentration of tokens that at least pass basic security checks.

That doesn't mean the market is "healthy." It means the garbage is temporarily taking a break.

Top 5 Picks of the Week

Here's what the algo surfaced, and what happened after.

1. $durr (durrcoin): 5.78x in 120 minutes

The week's biggest winner. Selected at about $67K market cap with 22% top-10 holder concentration. Hit its peak two hours later, then faded. If you weren't watching around 1 AM UTC on April 2, you missed the window entirely. Classic memecoin timing: the best move happened while most people were asleep.

2. $PEEPO (Peepo): 4.25x in ~24 hours

The slow burner. Selected at roughly $48K market cap with 24% top-10 concentration. Unlike $durr, this one took its time. Peaked after nearly 24 hours of tracking. The kind of token where patience actually got rewarded, which is rare.

3. $Buddy (Buddy Claude Companion): 3.72x in 12 minutes

Fast and gone. Selected at about $48K market cap but with a 31% top-10 concentration, pushing into warning territory on the wallet distribution. Hit 3.72x in twelve minutes, then reversed. The speed alone tells you: this was a coordinated move by early holders. If you got the alert and acted in seconds, great. If you hesitated for five minutes, you bought someone else's exit.

4. $Blicky (Blicky): 3.51x in 44 minutes

Mid-range timing, decent move. $67K market cap at selection, and the healthiest wallet distribution of the week at just 13% top-10 concentration. When the top 10 wallets hold less, there's less concentrated selling pressure. The data here matches the theory.

5. $Biscuit (Cryptococcosis Cat): 3.24x in 40 minutes

Selected at the highest market cap of the bunch, around $78K, with 18% top-10 holdings. Solid but not spectacular. About 40 minutes to peak, then the typical fade.

The pattern: four out of five top picks peaked in under two hours. The median hold time for memecoins dropped to 58 seconds according to recent data, and even the "winners" in our dataset reward attention spans measured in minutes, not hours. $PEEPO was the exception, not the rule.

Pump.fun Owns Everything (Still)

100% of this week's selections came from Pump.fun. All 32 tokens. Zero from Raydium LaunchLab (19 tokens tracked, 0 selected), zero from Meteora (4 tracked, 0 selected), zero from anywhere else.

This is the third straight week of total Pump.fun dominance on selections. At this point, if it didn't launch on Pump.fun, the algo basically isn't seeing anything worth flagging. Raydium LaunchLab keeps producing volume but nothing that passes the security and market health bar.

Pump.fun's average multiple for selected tokens: 2.02x. Not breathtaking. But it's 2.02x more than every other platform produced.

The Bundler Problem Isn't Going Away

Here's the number I'm watching most closely this week.

Average bundler concentration on selected tokens: 23.97%.

That's in dangerous territory. The threshold where bundler holdings start becoming a real concern is 15%. We're well above it. And these are the tokens that passed the filters. The ones that got through.

What this means: even among the "good" tokens, almost a quarter of the supply is held by wallets that bought in coordinated bundles before public trading started. These are insiders. They got in before you did, they got in cheaper than you did, and when they sell, the chart goes down.

Average dev holdings: 0.18% (healthy). Average insider concentration: 2.67% (warning zone). Snipers: 0.55% (fine). The bundlers are the outlier dragging the risk profile up.

This connects to last week's bonding speed analysis too. Faster bonding tokens tend to attract more bundler activity, and slower bonders tend to perform better. The data keeps reinforcing the same story: the early money isn't your friend.

One Story From the Data

The serial deployer stat of the week: a wallet behind the token $DOT had deployed 52,947 previous tokens. Not a typo. Almost fifty-three thousand tokens from one deployer.

$DOT made it through initial tracking but only hit 1.17x. A serial deployer at this scale is basically running a factory: deploy, see if anything sticks, move on. The algo flagged it as story-worthy not because it was a good pick, but because the sheer scale of deployment is a signal about what's happening at the bottom of the funnel.

For every token you see on your timeline, there are thousands of attempts like this happening in the background. Factories grinding through variations hoping one catches traction. 52,947 tries from one wallet. The market you see is a thin layer on top of industrial-scale token spam.

What This Means for Next Week

The selection rate tripling sounds good. And it is, kind of. More tokens passing filters means more potential opportunities per day.

But context matters. Volume is down. The macro environment is brutal. The Fear & Greed Index at 8 means the broader market is in survival mode. When fear is this extreme, liquidity dries up everywhere. A token can pass every security check and still go nowhere because there aren't enough buyers.

The algo doesn't predict markets. It vets tokens. It tells you which ones passed the security and supply checks. What happens after selection depends on the market showing up to trade.

Three things I'm watching:

• Whether the selection rate holds above 5% or reverts as new deployers re-enter
• Bundler concentration trends (if 24% becomes 30%, the "good" tokens start looking a lot less good)
• Any sign of volume recovery as we move deeper into April

The data says quality improved this week. The data also says almost nobody is trading. Both things can be true at the same time.

Key Takeaways

• 589 tokens tracked, 32 selected (5.43% selection rate, up from 1.89% last week)
• Median outcome: 1.74x for selected tokens. Best pick: $durr at 5.78x
• 87.1% hard rejection rate: the filter wall caught 513 tokens before they reached users
• Pump.fun produced 100% of selections for the third consecutive week
• Bundler concentration at 24% on selected tokens remains in dangerous territory
• Volume down 14% week over week in a macro environment of extreme fear

Data from Drill's algo, March 30 to April 6, 2026. All results include wins and losses. The Alpha Log shows everything.

Sources

1. Drill.meme proprietary data — blogData/daily, April 6, 2026
2. Solana Hits 10.1B Transactions in Q1 — CryptoTimes, April 4, 2026
3. Trump Tariffs, Drift Hack & Bitcoin Crash — SpazioCrypto, April 2026
4. Solana Price Outlook 2026 — OpenPR, April 2026
5. Solana Price Forms Bearish Flag — The Market Periodical, April 3, 2026