What if pre-1906 football was modernized: A reddit thread turned fake sport PC sim

[Young Drachma]

So this is a little different than my 20+ years ago dynasty at inventing an actual sport.

This time, it was fueled by a silly reddit thread that asked what if the 1894 Yale team came back to life?

I saw this at a random time and thought about it, mostly to wonder what the format would have to look like for them to be able to learn it and compete. After spec-ing out some rules, I wanted to actually see a box score of the sport. After seeing unrealistic results, I immediately thought "ok, i guess I need more."

So I started working on a simulator for myself for this very weird hybrid fake sport.

[Young Drachma]

Okay so here are the core rule for your newest favorite fake sport.

The Core Rules

The field and structure will feel familiar:

• 100-yard field, goal line to goal line
• 6 downs to gain 20 yards for a new set of downs
• Possession changes on turnovers, punts, scores, or failure to convert

But here's the big differentiator: There are no forward passes.

Instead, the ball moves through lateral chains. The ball carrier can pitch or lateral to any teammate behind them or beside them. That player can lateral again. And again. Each lateral in the chain increases the risk of a fumble, but also creates opportunities for explosive plays.

A typical play might have 2-3 laterals. An aggressive offense might chain together 4-5 in a single play. Every lateral is a calculated risk.

The Scoring System

Viperball uses a multi-tiered scoring system:

The Scoring System

Viperball uses a multi-tiered scoring system:

Touchdown - 9 points
Ball crosses the goal line

Snap Kick - 5 points
Drop kick through uprights (can be attempted on ANY down)

Field Goal - 3 points
Place kick through uprights

Safety - 2 points
Tackle ball carrier in their own end zone

Pindown - 1 point
CFL-style rouge — kick reaches end zone and isn't returned

Bell - ½ point
Recovering opponent's fumble or loose ball

That half-point scoring is one of Viperball's signature elements. Every fumble recovery matters. Final scores like 47½ to 44 or 52 to 51½ feel distinctly Viperball. You can win by half a point.

The fact that this isn't going to be a real sport is what allowed me the freedom to implement design choices that might be harder IRL.

Naturally, I'll do more than just explain here, we're gonna play test my engine through an actual dynasty.

[JonInMiddleGA]

Gives me Blood Bowl vibes (minus the death & dismemberment)

[Young Drachma]

Toolwise, I ended up using a combination of them to build out the spec. That wasn't really my plan initially, but since this was basically just gonna be a weekend project, I didn't want to spend a lot of time preening this thing and opted instead to just go right into adding the features I wanted to test and to create a relatively plausible boxscore/RNG and then I can go from there.

Like many of us, I've attempted to build sims in the past but the sheer complexity of the work relative to the interest people would have in playing it, given all of the other things online right now made that kind of tough.

LLM/AI buidling tools significantly decrease how much time it takes to spec something out, but it generates not great code and frankly, it can get super expensive very quickly if you're not comfortable auditing your own stuff and/or not paying for expensive subscriptions for other things already.

For this project, I already had access to Claude Code. The original idea actually started in a Copilot chat, before I moved over the GPT for some lightweight imagining of the box score and once that got going, I decided okay maybe somehing really basic could be okay... and before i knew it, I was using the last of my Replit credits to convert the python app for Viperball I'd built with Claude into some kind of basic UI using Streamlit.

So the workflow (this is a 4 day old experiment from writing the rules to building the tool) has been me serving basically as product owner between Claude Code acting as an engineer and quasi-PM and then delegating most tasks to Replit for the heavier stuff.

it's been an okay workflow, but like I've said this isn't the cheapest way to do this and if I hadn't already been using the tools it would be even less ideal. The other thing is, I kind of went heavy of big features almost from the outset because I knew I just wanted something a bit mature from the outside. When building for an audience of one, it's a lot easier to just optimize for the stuff you care about and not really concern yourself with marketability or fans.

Making a fake sport still has constraints

The easiest thing about building a text sim for a fake sport is that the sport doesn't exist. You can visualize it, but it's not real. So you're able to play G-d mode and essentially fiat things you want into existence that may or may not be realistic IRL, but fit your imagination of what should happen. (See 9 point touchdowns)

Being an alum of sports game designing (Toccer, 2004) I have a deep understanding of some of the motivations of sports design and I also have lots of critiques about why most modern sports are just layered onto existing ones because we're calcified in our incentive structures and it's hard to break out of.

With Viperball, it was an opportunity to take my best reddit ideas (tm) around leagues and realignment and couple that with my own sports takes about what makes a competitive game -- a sport where you're never really out of it, but where talent still can show itself -- and create mechanics that keeps fans engaged.

So many of the modern games that are most fun in-person like hockey or even baseball are terrible for television watching. Viperball is a sport you can't watch or experience because it's fake, so I wanted to devise ways for the box scores and stats to tell stories that you can almost imagine, especially since it's based on gridiron football.

I've long been obssessed with hybrid codes of football like the Aussie Rules/Gaelic merge, Finnish baseball (Pesäpallo which itself is a fork of modern baseball) and so the idea that you could create a hybrid version of gridiron that traces its roots and is an alternate history that it's made for women was enough bait for me to finish this.

Wait, a women's sports sim?

Yeah, sorry FM26 haters. I create a women's only text sim for college viperball. how's that for a niche? because I decided that Viperball would have an audience in a world where flag football is blowing up and tackle football for women hasn't captured the imagination. Rugby Sevens and even our football codes do well overseas, my northern European friends love the sport. So I could imagine a world where if you could make a fast and tactical version of game that feels closer to something they've seen and that you could actually play that it'd be kind of fun to model.


I'm trying to balance this dynasty with what's happening under the hood with actual gameplay from Viperball since honestly that's the only reason I bothered with this.

I'm also going to deploy this on a cloud server somewhere so folks can mess around with it, I know the UI is abysmal right now, but I only ever built this for myself to test (And it's come a long way in like 2 days) so I'm not going to invest in anything prettier yet until I feel like the thing is working better.

[Young Drachma]

Originally Posted by JonInMiddleGA

Gives me Blood Bowl vibes (minus the death & dismemberment)

amazing comparison

[Young Drachma]

Before I start this dynasty, you need to know who's actually out there. The positions in Viperball look different because the sport IS different. No forward pass means no quarterback in the traditional sense, no wide receivers, no tight ends. Instead you get this:


OFFENSE

Zeroback — This is your "quarterback" except she's not throwing anything. The Zeroback takes the snap, reads the defense, and decides: hand off, lateral, keep it, or kick. Yes, kick. The Zeroback is your primary snap kick threat. On any down, from anywhere inside about 55 yards, she can drop kick it through the uprights for 5 points. A great kicking Zeroback changes your entire offensive identity.

Viper — The playmaker. Every team has one, she wears #1, and she's the most dangerous player on the field. The Viper is a hybrid — she can run, receive laterals, and create chaos. Think of her as a weapon you deploy. In the Ghost formation, the defense literally doesn't know where the Viper is pre-snap.

Halfbacks — Your workhorses. Two to four of them on the field depending on your scheme. They take handoffs, they run power, they're the middle of your lateral chains. In a Ground & Pound offense these are your most important players. In a Chain Gang offense they're lateral relay stations.

Wingbacks — Flankers who line up wider. Speed and lateral skill matter most here. A Wingback with 90+ speed is a home run threat on sweeps. They're also your primary targets on extended lateral chains — the ball works from the interior OUT to the Wingbacks on the edge.

Shiftback — Motion specialist. Lines up in the backfield but shifts pre-snap. Creates misdirection, can take direct snaps in the Triple Threat offense. Think of the Shiftback as a chess piece — her value is in making the defense wrong.

Linemen — Five of them up front. They block on runs, they push the pile on power plays, and on defense they're your tacklers. A good offensive line doesn't show up in the box score but you feel it in yards per carry. On defense, their tackling and power ratings determine how often they blow up plays.

Wedge — Interior blockers who specialize in creating lanes for the run game and protecting the Zeroback on kick attempts. They're the unsung heroes. A bad Wedge unit means your snap kicks get blocked and your power runs go nowhere.


DEFENSE

Safeties — Deep defenders. They read the play, cover ground, and are your last line against breakaway runs. Awareness is their key stat — a Safety with high awareness jumps routes (lateral routes, but still) and closes on the ball carrier before the chain develops.

Corners — Edge defenders matched up against Wingbacks and Wing/Ends. Speed matters here because they're chasing the fastest players on the field. A Corner who can't run with a Wingback gives up huge plays on sweeps.

Wing/Ends — Defensive edge players. They contain sweeps, rush the Zeroback on kick attempts, and force the play back inside. Think of them as your chaos agents on defense.


PLAYER ARCHETYPES

Every player also gets an archetype based on her stats. A few examples:

- A Zeroback with 80+ kicking who kicks better than she runs? That's a Kicking Zeroback. She's your Boot Raid triggerman.
- A Viper with 90+ speed AND 90+ lateral skill? Receiving Viper. She's a nuclear weapon.
- A Halfback with 93+ speed? Speed Flanker. She breaks contain and goes 60 yards.
- A Safety with 90+ speed who's faster than she tackles? Return Specialist. She flips field position.

The archetype doesn't change what a player CAN do — everyone can lateral, everyone can run — but it tells you what she does BEST and how the engine weights her decision-making.

36 players per roster, I'm still working out the depth model in the game tbh.

[Young Drachma]

This is where Viperball gets its identity. Because there's no forward pass, every offense has to solve the same problem differently: how do you move the ball 20 yards in 6 downs when the defense knows you're either running it or kicking it?

The answer is that "running it" covers a LOT of ground in this sport. Here are the 10 offensive schemes in the sim:


GROUND & POUND
Old-school power football. Dive options, power runs, sweeps. 92% run preference. You're using all 6 downs to grind 20 yards. Your Halfbacks and Linemen better be good because you're running behind them 30+ times a game. Low variance, low risk, low lateral usage. You win 27-18 and the other coach hates it. If you've ever played a 1970s Big Ten team in a college football sim, you know this offense.

LATERAL SPREAD
The opposite philosophy. Stretch the defense horizontally with 2-4 lateral chains per play. The ball starts with the Zeroback, goes to a Halfback, who pitches to a Wingback, who laterals to the Viper cutting underneath. Every lateral increases fumble risk. But when it works, you get 15 yards because the defense can't cover the whole width of the field. High variance. You'll score 52 one week and 17 the next.

BOOT RAID
Air Raid but with feet. The entire offense is built around getting to "The Launch Pad" — opponent's 40-45 yard line — and then firing snap kicks for 5 points. Your Zeroback needs elite kicking. The offense is conservative between the 20s (run the ball, move the chains, don't turn it over) and then transforms once you're in range. A good Boot Raid team can hit 3-4 snap kicks per game. That's 15-20 points just from kicking, which frees up the run game because the defense is terrified of giving up field position.

BALL CONTROL
The most conservative scheme. You run the ball, you protect it, you take the 3-point place kick when you're in range, and you play defense. 95% run preference. Tempo is glacial — you're bleeding clock on every drive. This is the anti-fun offense that wins you games when your roster is mediocre and you can't afford turnovers. Win 24-21, shake hands, go home.

GHOST FORMATION
Controlled chaos. The Viper is the centerpiece — she touches the ball 35% of the time, which is absurd. Pre-snap motion on 80% of plays. Misdirection bonus of 1.3x. The defense doesn't know where the playmaker is, doesn't know who's getting the ball, doesn't know if it's a run or a lateral chain. If your Viper is elite, this offense is terrifying. If your Viper is mediocre, you're just confusing yourself.

ROUGE HUNT
Defense-first "offense." You punt early (sometimes on 3rd down), pin the opponent deep, and score on Pindowns, Bells, and Safeties. This is not a normal football philosophy. You are actively trying NOT to score touchdowns. Instead you want field position, forced fumbles, and your kick team pinning the ball in the end zone for 1-point Pindowns. A Rouge Hunt team might win 14½-12. It's ugly and beautiful at the same time.

CHAIN GANG
Maximum laterals, maximum chaos, maximum entertainment. Every play is a 4-5 lateral chain. 55% of plays are lateral spreads. Tempo is cranked to max. Risk tolerance is 90%. You will fumble. A lot. But when a 5-lateral chain breaks loose for a 40-yard touchdown, it's the most exciting play in sports. This is the scheme you pick when you want to have fun and don't care about your win total. Or when you have a roster full of players with 85+ lateral skill and you want to watch the world burn.

TRIPLE THREAT
Single-wing misdirection. Your Halfbacks take direct snaps 25% of the time. Power runs, speed options, sweeps — all from multiple ball carriers so the defense can never key on one player. Misdirection bonus of 1.2x. This is the thinking person's offense. Not as flashy as Ghost or Chain Gang but harder for the AI to read.

BALANCED
No strong tendency. Adapts to the situation. This is the "I don't know what I want to be yet" offense, or the "my roster doesn't fit any one scheme" offense. It's fine. It's never great. Pick this if you're a coward or if you genuinely have a roster that does everything okay and nothing exceptionally.


And on the other side, 5 defensive schemes:

BASE DEFENSE — Fundamentals. Does nothing poorly, nothing exceptionally. Your default until you know what you're facing.

PRESSURE DEFENSE — Aggressive blitzing. Destroys power runs and short-yardage plays but gets torched by draws and counters. High risk, high reward — you'll force more turnovers and also give up more big plays.

CONTAIN DEFENSE — Anti-chaos. Built to stop lateral chains and the Viper. If you're facing a Chain Gang or Ghost team, this is your answer. Lateral spreads get crushed (0.65x modifier). The tradeoff is you're softer against straight-ahead running.

RUN-STOP DEFENSE — Stack the box. Elite against the ground game. Dive options and power runs hit a wall. But you're giving up the edges — lateral spreads actually get a bonus against this defense. Don't run this against a Chain Gang team unless you enjoy pain.

COVERAGE DEFENSE — Anti-kick. Built to prevent Pindowns and shut down the punt return game. If you're facing a Rouge Hunt or Boot Raid team, this limits their kicking game. You sacrifice some run-stop ability for special teams dominance.

I'm still working these things out, and I'd love to see more innovation in the schemes, since we're just play testing or installs of customized playbooks off a set of core foundational sets like real football has, but I'm still trying to understand how this game would operate differently.

Part of why I started writing this here was trying to start posting box scores and seasons so that we can iterate on how the sport might adapt or evolved based on you all-knowing ball-knowers who have been simming and building playbooks for real football since the 90s in our fake football leagues.

[Young Drachma]

Alright, you know the positions and the playbooks. Now you need to know what the game actually looks like. If you've played football sims, a lot of this will feel familiar, but several things are fundamentally different and they change everything.

Here's the cheat sheet, then I'll break it all down:

FOOTBALL vs. VIPERBALL
──────────────────────────────────────
Forward pass? YES NO
Laterals? Rare THE WHOLE OFFENSE
Downs to convert 4 6
Yards for first down 10 20
Touchdown 6 pts 9 pts
Extra point attempt Yes No
Field Goal 3 pts 3 pts
Drop Kick Extinct 5 pts (the Snap Kick)
Safety 2 pts 2 pts
Rouge/Pindown N/A 1 pt
Turnover Bonus N/A ½ pt (the Bell)
Overtime Yes No (ties stand)

I know. That's a lot of differences. Let me walk through each one.


═══════════════════════════════════════
SIX DOWNS, TWENTY YARDS
═══════════════════════════════════════

This is the single biggest adjustment if you're coming from football.

You get 6 downs to gain 20 yards. Not 4 downs for 10. The math works out differently than you'd think. In football, you need 2.5 yards per down to stay on schedule. In Viperball, you need 3.33. That's a harder ask per play, BUT you get two extra chances, and no defense knows when you're going to kick because you have so many downs to work with.

The extra downs also change decision-making completely. In football, 3rd down is the critical conversion down. In Viperball, there's no single "money down" — you've got downs 4, 5, and 6 all as viable conversion opportunities. Teams that play conservatively on down 4 and go aggressive on 5 behave differently than teams that attack from down 1 and treat down 4 like football's 3rd down.

If you don't convert by down 6, it's turnover on downs. Same concept as football, just later.


═══════════════════════════════════════
NO FORWARD PASS. ONLY LATERALS.
═══════════════════════════════════════

There is no forward pass in Viperball. The ball only moves forward by running it. Laterals — backward or sideways tosses — are legal and are the primary way the offense creates mismatches and moves the ball to the edge.

This means the Lateral Spread play family is the closest thing to a passing game. The ball gets handed off or tossed sideways in chains of 2-5 laterals, moving from player to player across the field. Each lateral in the chain adds fumble risk (roughly 3.5% per exchange). A 2-lateral chain is relatively safe. A 5-lateral chain across the entire width of the field is a dice roll — but if it works, it can break a game open.

If a lateral goes forward instead of backward, that's an Illegal Forward Lateral — 5 yards AND loss of down. One of the most punishing penalties in the game because you lose the down AND the yardage.


═══════════════════════════════════════
SCORING
═══════════════════════════════════════

This is where Viperball gets weird. Good weird.

TOUCHDOWN — 9 POINTS
Get the ball to the end zone. Same concept as football. But it's worth 9 points, not 6, and there is no extra point attempt. You score 9, you kick off, that's it. The higher value means every TD is a massive swing. A 9-0 lead feels like 14-0 in football.


FIELD GOAL (PLACE KICK) — 3 POINTS
Same as football. Your Zeroback lines up, the ball is snapped, she kicks through the uprights.

SNAP KICK (DROP KICK) — 5 POINTS
This is the uniquely Viperball scoring play. The Zeroback catches the snap and immediately drop-kicks it through the uprights. No holder, no setup — she catches it and kicks it in one motion. It's worth 5 points because it's significantly harder to execute.

Mostly I was trying to picture a fork of football, and imagined that the kickers would have kept the game going, but I'm thinking that in an era where flag is dominant that you could get Aussie exports, Rugby kids, and former soccer players to make this position really shine.

The beautiful thing about the snap kick is that it can happen on ANY down, not just kicking situations. There's roughly a 12-18% chance per play that your ZB attempts one (higher if she has a "Kicking ZB" archetype). It's a surprise weapon. You're on 2nd & 14 and instead of running another play, your ZB just drop-kicks it from 38 yards. If she hits it, that's 5 points out of nowhere.

The risk: the opposing Keeper can deflect a snap kick. About 8% base chance, higher for athletic Keepers. If the Keeper deflects it, the kicking team loses possession AND the defense gets a ½-point Bell. So a missed snap kick isn't just zero points — it actively hurts you.

Elite kicking ZBs hit snap kicks at 94% from 25 yards. Bad ones hit at 64%. From 55+, even elites are in the 40s. It's a high-variance play that rewards having a genuinely good kicker.

SAFETY — 2 POINTS
Ball carrier gets tackled in her own end zone. Same as football. Probability scales by field position — 14% chance inside the 2-yard line, dropping to about 2.5% at the 15. Scoring team gets the ball back.

PINDOWN — 1 POINT
This is Viperball's version of the rouge from Canadian football. When a kick (punt, field goal attempt, or snap kick) lands in or through the end zone and the receiving team can't return it out, the kicking team gets 1 point.

The receiving team gets a chance to run it out. Their chance scales with team speed. A fast return team can field it and get out of the end zone. A slow team gets pinned down — hence the name.

One point doesn't sound like much until the game is 27-27 in the fourth quarter and a well-placed punt earns you a pindown to take a 28-27 lead. It adds this entire layer of strategic kicking that doesn't exist in American football.

THE BELL — ½ POINT
When the defense recovers a fumble, blocks a kick and recovers it, or forces a muffed punt and recovers, they get a half point in addition to the possession change. It's called a Bell because the original scoreboard mechanic was a literal bell that rang in the stadium.

Mostly I thought it was weird that no sport truly has fractional points besides judged events.

═══════════════════════════════════════
HOW A DRIVE ACTUALLY WORKS
═══════════════════════════════════════

Here's a typical possession:

1st & 20 at the 25. POWER — gain of 6.
2nd & 14 at the 31. SWEEP OPTION — gain of 8.
3rd & 6 at the 39. SPEED OPTION — bad read, loss of 2.
4th & 8 at the 37. DIVE OPTION — gain of 5.
5th & 3 at the 42. COUNTER — defense bites on misdirection, gain of 11. FIRST DOWN.
1st & 20 at the 53. ...new set of downs, keep driving.

Each play, the engine:
1. Selects a play family based on your offensive style, down, distance, and field position
2. Rolls base yardage (gaussian distribution around the play family's expected gain)
3. Applies modifiers — Viper alignment, defensive scheme matchup, player archetypes, fatigue, weather, game rhythm
4. Checks for explosive plays (big gains that break through the secondary)
5. Checks for turnovers (fumbles, with probability modified by play type, weather, and fatigue)
6. Checks for penalties

The result is a yards gained number that gets applied to field position. If you've gained enough for a first down, your downs reset. If not, next down.

Each play eats 15-45 seconds of game clock, modified by your team's tempo. A high-tempo team burns through the clock faster because they run more plays per quarter. A slow, grinding Ground Pound offense eats clock on every snap.

Drives end when you score, turn the ball over, punt, or time runs out. There's a maximum play cap per drive (scales with tempo) to prevent infinite possession loops, but you'll almost never hit it in normal play.


═══════════════════════════════════════
TURNOVERS
═══════════════════════════════════════

No interceptions in Viperball. No forward pass means no picks. Turnovers come from three places:

FUMBLES ON RUNS
Every play has a fumble probability based on the play family. Conservative plays are safer:
Dive Option: ~1.0% fumble rate
Power: ~1.2%
Counter: ~1.5%
Draw: ~1.4%
Sweep Option: ~1.6%
Speed Option: ~1.8%
Viper Jet: ~2.8% (HIGH risk, high reward)

These rates get modified by weather (rain adds 2.5%, sleet adds 3.5%), player ball security, defensive pressure, and fatigue. A tired team running Viper Jets in the rain is begging to cough it up.

When a fumble happens, there's a recovery roll. Defense recovers roughly half the time. If defense recovers, they get the ball AND a ½-point Bell. If the offense falls on it, they keep possession but advance the down.

LATERAL CHAIN FUMBLES
Lateral Spread plays chain the ball through 2-5 players. Each exchange carries ~3.5% fumble risk. The math gets ugly fast:
2 laterals: ~7% cumulative risk
3 laterals: ~10%
4 laterals: ~13%
5 laterals: ~16%+

That 5-lateral chain is a one-in-six chance of putting the ball on the ground. The upside is massive yardage if it works. The downside is a turnover, a Bell, and your coach screaming.

TURNOVER ON DOWNS
Fail to gain 20 yards in 6 downs and you hand the ball over. No points for the other team, but field position flips and they start fresh with 1st & 20.

═══════════════════════════════════════
PENALTIES
═══════════════════════════════════════

The game has 48 penalties across five phases. I'm not going to list all 48, but here's what you need to know:

PRE-SNAP (before the play happens):
These kill the play entirely. False Start and Offsides are the most common (5 yards each). Delay of Game, Illegal Formation, Encroachment — all 5 yards. The Viperball-specific one is Illegal Viper Alignment — your Viper lined up where she wasn't supposed to. 5 yards.

DURING THE PLAY:
This is where the big ones live. Holding is 10 yards on offense and happens more than anything else. Defensive penalties like Facemask (15 yards) I don't think we'd have actual helmets in this sport but maybe field hockey face guards and it'd still count, Horse Collar (15), and Defensive Holding (5) all come with automatic first downs. That's huge in Viperball because a first down resets you to 20 yards to go — an automatic first down on 5th & 12 is a massive gift.

Lateral chain plays have their own penalties: Lateral Interference (defense, 10 yards, auto first down), Illegal Forward Lateral (offense, 5 yards PLUS loss of down — devastating), and the Viperball-specific Illegal Viper Contact (defense, 10 yards, auto first).

KICKING PLAYS:
Roughing the Kicker (15 yards, auto first) is the big one. Running Into the Kicker is only 5 yards without the automatic first down. Fair Catch Interference and Kick Catch Interference are both 15-yarders.

AFTER THE PLAY:
Taunting, Unsportsmanlike Conduct, Late Hit, Excessive Celebration — all 15 yards. These trigger after the play resolves and can flip momentum even after a big gain. Score a 40-yard touchdown? Your Wingback's Excessive Celebration just moved the kickoff back 15 yards.

PENALTY INTELLIGENCE:
The AI knows when to decline penalties. If the defense forces a fumble, they decline the holding call — the turnover is worth more. If the offense breaks a 30-yard run, they decline Defensive Holding — the yardage gained exceeds the penalty yardage. This matters because it means the engine isn't just rolling dice; it's making the strategically correct call.

ENVIRONMENTAL FACTORS:
Bad weather increases penalty probability by 0.002-0.003 per penalty. Late-game pressure (4th quarter, under 5 minutes) multiplies all penalty rates by 1.15. Teams get sloppier when it matters most.

═══════════════════════════════════════
SPECIAL TEAMS
═══════════════════════════════════════

PUNTS:
Used when you're too far for a kick and don't want to risk going for it. Distance varies by field position and kicker skill. The receiving team has to field it — and about 25-30% of the time, they muff the catch. If the punting team recovers the muff, they get a Bell. There's also a 4% chance of a tipped punt (ball stays live at the line — chaos) and a 7% chance of a weird bounce that moves the landing spot 10-25 yards in either direction.

Punts can also result in a Pindown if the ball reaches the end zone and the returner can't get it out. This makes field position punting — kicking it deep with the intent to pin them inside the 5 or earn a Pindown point — a legitimate late-game strategy.

This frankly is the hardest part of the game to work on because when I first started building this, the games were absurdly low scoring punt-fests because it was basically making a rugby scrumfest. I've worked on getting things to where it feels more dynamic and the scoring is higher as a result. Not Aussie rules basketball on grass scoring, but...far higher than your garden variety football game.


KICKOFFS:
Standard kickoff after scores. Distance ranges 40-85 yards with returns of 10-30. Long kicks (60+ yards) risk giving up a Pindown point but pin the opponent deep. Short kicks are safer but give better field position.

BLOCKED KICKS:
Place kick blocks happen 5-15% of the time. Punt blocks are 4-8%. Blocked kicks create live-ball situations where either team can recover. If the defense scoops and scores, that's 9 points plus a Bell. 9½ points off a blocked kick. It happens.

═══════════════════════════════════════
THE CLOCK & GAME FLOW
═══════════════════════════════════════

Four 15-minute quarters. 3,600 total seconds. Each play burns 15-45 seconds of clock, modified by team tempo.

If the score is tied after four quarters, the game ends in a tie. No overtime. This is intentional.


═══════════════════════════════════════
WEATHER
═══════════════════════════════════════

Wind is the kicking disruptor. Your field goals and snap kicks lose accuracy, but your punts travel farther. A windy game rewards teams that can score on the ground and use punting as a weapon. Heat, cold and sleet also have chaos factors.

That's the game. Nine types of points. Six downs. No forward pass. Fumbles are your interceptions. The Viper and the Keeper create chess-match alignment decisions every single play. And the weather can turn any gameplan upside down.

Next up, the conferences and teams. I mostly opted for D1 programs without football and random others.

There's still a lot of tweaking to do, but I think what's gonna happen now is after my last set of tweaks I'm just gonna start trying to run a season or 2 and see how it goes.

[Young Drachma]

We've got 186 teams in this league across 16 conferences and every state and Canada are represented.

My naming engine lets international players in, we've got a crude NIL implementation and a transfer portal too.

I think what I'm gonna do now is play test the season mode, then we'll play test dynasty mode and see where I find the bugs, see how it feels and how much the coach can impact things, because right now it's just a simulator not really an engine where you can impact a whole lot besides recruiting (in dynasty mode) in season mode you're just watching after setting your offensive/defensive scheme from the outset.

So there's not a lot of variety yet, which I might try to work on so that it's not too predictable, but I also don't want to introduce even more complexity if it's not helping or enhancing gameplay.

[Young Drachma]

Because I thought the old school dynasty format was kind of boring and because I'm me, I built a fantasy sports/gambling engine atop Viperball and we'll be sharing results in part from what that generates in season mode or eventually in dynasty mode.

DraftyQueenz is a self-contained prediction/fantasy/currency system that runs alongside each Viperball season. It adds a metagame layer where the user manages a DQ$ bankroll through gambling and fantasy, then funnels winnings into dynasty program boosts.

I wanted a way to make God mode more fun than just doing what I normally do and to have some real randomness involved. Since I know nothing about this universe, the teams, the players and it's all random each time anyway, it'll be a worthwhile exercise.

At some point, the dynasty mode will be better built out, but I can't really know that until I spend some time actually grinding out a few seasons within it.

[Young Drachma]

Some of the things with building this game quickly and sort of iteratively is that stuff I originally built that I revamped were still legacy baked into the code. For instance, injuries were cosmetic until I fixed that last night and implemented a new injury/depth chart/fatigue model, so that there's actually some implication to players grind.

The other thing I removed were some early engine boosts to improve the run game -- after all, nothing says you can't still run the ball traditionally -- but initial test runs of the game were only focused on lateral chains, which wasn't really what I was hoping for, so I improved the run model.

The other new innovations isn't technically against the rules, it just hadn't occured to me -- kick passing. Think Gaelic football not Aussie rules marks -- but it would truly open up the field if players can still catch passes but they don't have to be thrown.

I finally feel like the engine is where I want it to start playing some seasons. The UI is in Streamlit and that's very ugly, but given this is a tool that's just for me, I didn't really want to invest in something more substantial.

It'll work great for dynasty reports though lol.

[Young Drachma]

So today, I ran a bunch of tests as I've been doing each day of working on this and trying to refine the engine and fix some of stock assumptions baked into the original game.

Stat	Current	Target	Status
TDs/team	2.62	3-5	A touch low but in the neighborhood
FGs/team	2.15	3-5	Low — teams aren't attempting enough
DKs/team	2.27	3.5-4	Low — same issue
Fumbles/team	1.16	~1.5	Close, moved in the right direction
KP Comp%	63.5%	55-70%	Right in the sweet spot
Yards/team	519	—	Healthy
Score/team	42	—	Reasonable
Punts/team	1.59	—	Low (good — teams are going for it)

The biggest issue with this game/sport when I first designed the rules was the whole "lateral chains" offense thing. I could picture it, after all, who hasn't watched some rugby but I thought it was going to be pretty ugly and not really a sport I personally wanted to watch. But it's what we started with.

Eventually, I started adding the offensive methods like drop kick points, opening up the run game more in the engine and eventually -- last night -- got the realization that the rules don't prevent kick passes. I had always imagined them, I just didn't put them into the rules or into the game at all.

So last night, I built that functionality into the game but because it wasn't really part of the core game or the playbooks, its effectiveness was extremely limited. Teams were using lineman to kick pass instead of the zeroback/viper combos I was picturing during rules development.

So today's task was rewriting the offensive engine from scratch basically to remove a lot of the random number generation (RNG) that was baked into the game from the start and to let the engine operate without constraints on outcomes on the field. Still, this didn't really address the ways offenses work and more importantly, defenses are still mostly cosmetic at this stage.

I'm finally working on another update that'll dramatically create new defensive styles. I want it to be possible to play and win with average players and a good defensive scheme. Down the road, the trick to this game will be the ways you can develop and devise schemes.


I've been describing Viperball as Go, with Gridiron Football as Chess.

When you're building for an audience of one, you're both always done and never done. But I think that so long as there aren't any showstopping bugs (and there were yesterday, so I didn't sim as much) we'll be able to run through a season.

Rethinking the kickoff

the balance between "don't make shit weirder" and "it's an original sport" is tough.

Someone on a NFL thread on reddit had this idea that I decided to poach for viperball for now:

Possession always alternates, but you start at the 20+/- however many points you are ahead or behind. So if Team A gets the ball down 14, they start on the 34. If team B gets the ball up 21, they start on their own 1 yard line.

This feels a little "unfair" but from a sports design perspective I like it. The game doesn't need to get "harder" but it should make it fairer because not all scoring is fair. Building essentially a handicap into the sport without giving away points is the most interesting way to make the game interesting in a realistic and plausible way (you could kick your way to the same outcome) but with more strategic stakes.

The yards assessed here are called "sacrifice yards" and they get measured by team annually and it's generated some fun efficiency statistics that i'm excited to dig into as well.

Looking forward to actually playing.

The latest game tuning was producing wildly out of whack numbers, the underlying engine was making a lot of assumptions about yardages and you had teams generating tons of yards and barely converting on them into touchdowns. For the entirety of this game so far, touchdowns have been too rare and my goal has to been to fix the engine to make them work better.


I added a lot of modifiers for different plays that were compounding yardage but not scoring actual touchdowns, but it was a good exercise in understanding both code and model complexity. This stuff can get bloated very quickly, and you need to be able to track what you're aiming for or else it can get away from you.

I've made the necessary adjustments, so we'll see how much closer we can get to actually deploying a season.

[SirBlurton]

This is completely unhinged in the best possible way. I love it, I’m extremely curious about your process for designing the engine.

I was an “early adopter” on using coding support and then got into AI music generation for a bit when the frustration got too much. I was using Antigravity last night and it’s really good - dusted off some prototypes and had fun.

One prototype is blood bowl, have programmed a Gridzone app, so this is right up my interests alley! Will be watching with interest!

[Young Drachma]

Originally Posted by SirBlurton

This is completely unhinged in the best possible way. I love it, I’m extremely curious about your process for designing the engine.

I was an “early adopter” on using coding support and then got into AI music generation for a bit when the frustration got too much. I was using Antigravity last night and it’s really good - dusted off some prototypes and had fun.

One prototype is blood bowl, have programmed a Gridzone app, so this is right up my interests alley! Will be watching with interest!

Thanksss. Yeah, it's for sure the most unhinged project in a long time haha. So I've basically got a crude AI engineering setup where Claude Code acts as my lead product engineer. Since I was already using it for something else, I've been using Pro and trying to maximize what I have left because i'm downgrading it at the end of the month (true of all of the tools I'm using) so I need to get it all done now while i have increased limits.

The app is built in Python and I've got a Streamlit UI so it's very ugly, but I sort of didn't know I was going to end up doing this much with it or else I'd have started with a React or some other kind of frontend UI rom the jump. It's still fairly easy to switch, I just am not going to be other unless this is something i end up deciding to ship publicly (besides here at FOFC)

So basically like I said, Claude Code is the lead engineer and then I've been using Replit for everything else and I've been effectively product owner, QA, strategist, IA and everything in-between. Because the repo is on github, it's made stuff relatively easy to work between them. Having Claude do the heavy lifting on some of the aspect of strategy and planning or even making it write specs of stuff it does on the backend, before Replit can connect it to the Streamlit API has kind of been a decent workflow.

I sort of backdoored my way into this workflow, weirdly the idea sort of sprouted when I was using GPT or something to spec out the original box score and concept when I realized "oh I have to build this, don't I?"

So yeah that's the basic gist of the engineering stack. Without pro limits in both of those tools, there's no way I'd be this far along. Like I said, if I ever wanted a better UI like even to the FOF level, but I've been working harder at testing the actual game to make sure it feels plausible and works first before making it prettier. I think it was a good decision to operate this way.

[Young Drachma]

So tonight's iterations happened while on a plane.

I did some simulating and found some issues with the lack of touchdowns, the game was generating far too many yards due to multipliers in the engine were stacking.

A lot of this stuff originated from my original sim, when I was just looking for a plausible box score and nothing else, but now with more of an actual game happening, I wanted most of that gone.

TDs were barely happening and that didn't make sense to me, given their point value, so I reengineering it to be more important followed by snapkicks, field goals or worse, punting. But the TD thng -- along with increasing the amount of plays in a game was too much and it become Aussie Rules style basketball scoring on grass.

The app now uses a NiceGUI UI instead of Streamlit, my goal is to connect more of teh dots, get scoring balance better but changing the fundamental ways that scoring was being done.

I actually got both engineers to give me summaries of how this has all come together. I'll post those next, because it's evolving very quickly. Showing the rapid scale is kind of both to show how much it's changing but also to help you understand under the hood how my assumptions have evolved over the course of the week.

The real thing is that none of this would be possible with a sport anyone has actually played, because there'd be a mental model for how it looks and lots of constraints on the design. With this, there's none of that and it's the only freeing thing about building this game like this is that I'm just trying to make something legible to me and that I want to play.

It went from being "what's a silly diversion" to "can I run decades in this and get immersed?"

[Passacaglia]

Seems awesome, in for one!

I'm curious about it being a women-only sim. Is there anything that makes it different from if it were men-only?

[Young Drachma]

Originally Posted by Passacaglia

Seems awesome, in for one!

I'm curious about it being a women-only sim. Is there anything that makes it different from if it were men-only?

No, the programming doesn't factor weights or heights so I could easily swap the name sets into mens viperball and it would play exactly the same way. It was just another quirk I was thinking about.

I guess at some point, I'll go add some dude names from NCAA and other sites and use the same teams and generate men's teams, I hadn't thought about it. Hah.

[Young Drachma]

This experience has been way more about 1) game design and 2) game mechanics and I guess broadly 3) design of a new sport than I was anticipating off a silly reddit post.

For context, I've made a lot of iterations of the game since the start

Initially I used Streamlit Cloud for the game UI because it was easier to spin up

Right now, the game operates in a NiceGUI interface that's a lot better to look at, and even this UI has had some improvements in the past few days, but it still is left with more improvements still to make -- it's not the greatest game on a mobile device -- but as a simulator only tool, it's gotten more efficient.

Part of doing this -- for myself -- as quickly as I have, is I'm not even sure what the narrative arc should be or how to explain it all.

Having simulated at this point thousands of games to get a sense of the game, the mechanics and feel of the sport. I've been working more to do more than just generate boxscores, but it's led me to work on other aspects of the tool and thinking more about the sport's philosophy.


Gone from a math sim to a proper game

The first iterations of this too were essentially optimized to do very simple things, which is make the box score look plausible. The difficult in pulling that off and frankly, not understanding what I was trying to make with the sport initially -- the lateral sport I first invented based on the reddit post, was very boring to me -- made me start thinking about what to do next.

So the first iterations of the simulation were extremely heavy handed in terms of dictating plays, distances and what results would happen basically a super dice roller. Ratings were cosmetic, teams themselves were basically cosmetic , decisions didn't have a real connection to anything happening on the field.

That's now changed through expanding the ratings, where to the dice rolls connect to the ratings to juice them or diminish, and the offensive and defensive play styles were built out substantially.

Still, a lot of the v1 code still powers the game and it's not until my play testng that I'm able to start to decipher where things don't feel right/good to me and what I'm wanting to do about any of it. That started with isolating the player positions themselves, then figuring out if there were schemes and how the game flow worked in this sport.

Slowly but surely, that's been evolving. Kick passes, for instance are new and the playbooks barely use it. The game is still emphasizing the run, even as i've essentially phased out laterals as the primary mechanic of scoring in this sport.

Because I can batch sim lots of games and run data on what's working, it's essentially like simulating decades of games at a time, and lets me build more rapidly than I think I'd otherwise be able to.

One of the other things I've been doing is writing reports on changelogs and data changes partially so I get a sense of how much is evolving and when I did it, because I'd find it difficult to keep track otherwise.

[Young Drachma]

I build a stats site that'll let me basically report out data, box scores & such from the simulation. The next level would be some future kind of hattrick interface but for now, I'm more interested in shareouts and building the storylines out.

There's men's Viperball now thanks to Passcaglia's question, but in the form of a pro leagues that don't interact with college & an international sim that does draw from the college game entirely.

When I'm on a proper computer, I'll share more.

[Young Drachma]

Did a huge overhaul where I replaced a bunch of the smaller schools for actual D1 FBS style programs, having decided I'm ready to see what that looks like in the box scores, it'll feel more like college basketball with small and big schools alike.

[Young Drachma]

DELTA YARDS
I'm not sure where Delta Yards came from. It was not my original idea, but I think someone on reddit suggested that we should replace kickoffs with this system that would move the ball +/- your own 20 yard line based on how much you were winning or losing. In trying to decide what to do with kickoffs in viperball, I decided to test it out and see how it works. The end result? Delta Yards. Viperball comes with its own philosophy around protecting leads called Lead Management, which takes your 2 minute drill and builds far more rigor around how teams shape their play to score points but also defenses to protect leads.

i've read that the "rubber banding" mechanics of viperball are fundamentally unfair -- no other sport would penalize a team that's winning with worse field position -- but I was curious to see how much or how often teams with leads blow them due to the inherent "unfairness" of the delta system. So the box scores for this sport track the yards we give away to trailing teams, as well as those by teams that have leads. Having run at this point thousands and thousands of games, it appears that most of the time teams that were going to win do so regardless.*

Delta system does inflate the yardages in games but again, I think having a mechanic that keeps the game potentially interesting for all 4 quarters is probably another unique component of this sport. Do you *want* your opponent to score early so you can take their early lead and use it for slightly better field position? All of this could happen in a world with kickoffs anyway, we just cut to the chase and save time by essentially dictating where you go based on what's going on.

The reason I've kept this system into the game was really around the imagination. Lead Management, imagining the call sheets for leads and trailing at various field positions based on where you are in the game, it adds a depth and complexity to Viperball and more strategic opportunities than football which already comes with a kind of complexity and rigor that fans around the world appreciate.

I used to call Viperball "Go" to Football's "Chess" but a more accurate representation would be "Backgammon with Mario Kart game mechanics"

[Young Drachma]

One of the weird rules of this Viperball thing is that interceptions reward you with a bonus possession. It feels like overkill, but it's part of the mechanics and I can't even decide if I like it, but the engine was rewarding the team who threw the pick with a possession after losing the ball which was not great.

It's good that I hadn't shared any of the stats yet, it'd all would've been wrong. In general, it didn't seem to be giving teams advantages that hadn't otherwise earned them.

I'm going to run a mock season, it'll be interesting now that we have actual major universities in the universe, and we'll report back with some interesting data points, ways this sport is different and stuff I'm still debating how much I like or dislike how things are coming together.

Other thing I can't figure out is the voice for this sport. It's not quite football, so it's been hard for me to figure out how to talk/write about what's happening in a game with game logs/box score, but on the flip side, the sport feels more plausible than it did a month ago when I started writing about it.

[Young Drachma]

Nebraska Omaha 74.5, Arizona State 72

ALBUQUERQUE — With three minutes left and her team up two, Veera Keto took a pitch on fourth-and-7, cut outside, and ran 60 yards for the touchdown. The 50,403 at FisherPrice Stadium — split roughly between Arizona State red and Nebraska Omaha crimson, with the Sun Devils faithful having made the short drive and the Mavericks' fans having made something considerably longer — took a collective breath, and then the Nebraska Omaha section erupted.

"We just kept running the ball," Keto said afterward. "That's all we know how to do."

Nebraska Omaha won the national championship 74.5–72, completing one of the more unlikely postseason runs in CVL history. The Mavericks entered as a 17-seed. Nobody here predicted this.

Arizona State was the better team on paper — higher rating, more yards on the night, Da'Brya Hydrick going 17-for-22 for 377 yards and two touchdowns in what may have been the finest individual performance on a losing side this season. Her team scored 72 points. It wasn't enough.

"I don't know what else she could have done," Arizona State coach Charles Perez said. "That's on us, not on her."

The game was played in heavy wind and it showed — seven scores in the first quarter, six lead changes total, both teams turning it over twice before halftime and scoring anyway. Sabou Mauigoa made four drop kicks through the gusts. Mindy Pope ran for two touchdowns on Viper jets, which nobody predicted. Julia Keeley separated her shoulder in the third quarter and didn't return. Nebraska Omaha kept running.

"Julia goes down and nobody blinks," Thomas Jackson said. "That's what this team is."

Arizona State tied it. Took the lead. Mauigoa answered both times. Keto's run made it a two-score game and then Arizona State drove to midfield in the final minute — three incompletions, a field goal from 47 yards to make it 74.5–72 with no time left. The FisherPrice scoreboard held the final score for a long moment before the Nebraska Omaha bench cleared.

Jackson has been coaching for 22 years. He is 217-101 lifetime. He had never won a national title before tonight.

"A lot of guts," he said. "That's what it took."

CVL NATIONAL CHAMPIONSHIP
FisherPrice Stadium · Albuquerque, NM · Heavy Wind

        1      2      3      4      F
ASU    31.5   18.0   13.0    9.5   72.0
UNO    27.0   21.5   15.0   11.0   74.5

TEAM STATS
                          ASU        UNO
Total Yards               666        527
Plays                      65         55
Touchdowns                  5          6
Rushing              289 (30)   373 (36)
Yards/Carry               9.6       10.4
Kick-Pass        16/22, 377y   7/11, 154y
KP Touchdowns               2          2
KP Interceptions            0          1
Drop Kicks                1/1        4/4
Place Kicks               4/4        0/1
Turnover Margin            -1         +1
Fumbles Lost                2          1
Penalties          10 (120y)    4 (25y)
Stop Rate               38.5%      14.3%
Kill Rate          9/9 (100%)   0/1 (0%)
Bonus Possessions           1          0

ARIZONA STATE — OFFENSE
Name                Pos  TCH  YDS  TD  Notes
Da'Brya Hydrick      ZB   22  377   2  17/22, 77% KP, 0 INT
Hunter Barrett       WB   11  261   1  110 rush, 151 rec
Ursula Bakker        HB   14  189   1  108 rush, 81 rec (inj. Q3)
Jessica Ricks        ZB    4   84   0  1/1 DK, 4/4 PK
Mackenzie Williams   HB    9   51   2
Victoria Rios        HB    1   46   1

ARIZONA STATE — DEFENSE
Name               Pos  TKL  TFL  SCK  HUR  INT
Kimberly Colvin     KP   13    0    1    0    0
Oona Paavilainen    DL   11    0    0    0    1  (INT returned for TD, Q1)
Tiare Wiki          DL    9    0    0    0    0
Oona Wijk           KP    8    0    0    1    0
Bethany Gaines      DL    5    1    0    0    0
Hanna Flores        DL    2    1    0    0    0
Aurora Wijk         KP    1    0    0    0    0

NEBRASKA OMAHA — OFFENSE
Name                Pos  TCH  YDS  TD  Notes
Lucie Pelletier      ZB   11  154   2  7/11, 64% KP, 1 INT
Veera Keto           SB   12  118   1  Winning TD — 60-yd run (Q4)
Julia Keeley         WB   11   95   0  Shoulder sep. Q3
Mindy Pope           VP    3   77   2  Both TDs on Viper Jet
Emily Rodriguez      WB    5   77   1  Sub for Keeley
Hanna Durfee         WB    2   81   0
Vicki Pitchford      VP    2   38   1
Sabou Mauigoa        ZB    —    —   —  4/4 Drop Kicks

NEBRASKA OMAHA — DEFENSE
Name               Pos  TKL  TFL  SCK  HUR  INT
Emilia Schmitt      DL   10    0    1    0    0
Rashida Robinson    DL   10    1    0    0    0
Nadia Sama          KP    9    1    0    0    0
Larissa Souza       KP    8    0    0    1    0
Heather Kiesewetter DL    7    0    0    1    0
Te-Hina Cole        KP    3    0    0    0    0
Mia Harthan         DL    3    0    0    0    0
Vilde Kemppi        DL    3    0    0    0    0
Dionne Spencer      DL    2    0    0    0    0
Tara Lynch          DL    2    0    0    0    0

DRIVE SUMMARY
 #  Qtr  Team  Start  Plays  Yds  Result
 1   Q1   ASU     20      3   65  Touchdown
 2   Q1   UNO     29      1    0  Int Return TD — Paavilainen 39 yds (-9yd)
 3   Q1   UNO     38      5   72  Touchdown (-18yd)
 4   Q1   ASU     20      2   65  Touchdown BONUS (+9yd)
 5   Q1   UNO     38      3    8  Fumble — Bell (-18yd)
 6   Q1   ASU     54      1    0  Punt
 7   Q1   UNO     21      1    8  Fumble — Bell
 8   Q1   ASU     71      5   67  Touchdown
 9   Q1   ASU     10      7   70  Field Goal (+10yd)
10   Q2   UNO     33     11   72  Touchdown (-13yd)
11   Q2   ASU     16      3   88  Touchdown (+4yd)
12   Q2   UNO     33      1    0  Punt (-13yd)
13   Q2   ASU     44      2    9  Fumble — Bell
14   Q2   UNO     47      2   53  Touchdown
15   Q2   ASU     16      6   95  Touchdown (+4yd)
16   Q2   UNO     32      5   26  Snap Kick (-12yd)
17   Q3   ASU     12      5   64  Snap Kick (+8yd)
18   Q3   UNO     32      3   57  Snap Kick (-12yd)
19   Q3   ASU     12      5   14  Snap Kick (+8yd)
20   Q3   UNO     30      5   39  Snap Kick (-10yd)
21   Q3   ASU     14      8   86  Touchdown (+6yd)
22   Q3   UNO     34      2   17  Stall (-14yd)
23   Q4   UNO     51      4    3  Snap Kick
24   Q4   ASU     10     10  101  Field Goal (+10yd)
25   Q4   UNO     32      5   73  TOUCHDOWN — Keto 60-yd run (-12yd)
26   Q4   ASU     16      5   47  Touchdown (+4yd)
27   Q4   ASU     25      9   38  Field Goal (-5yd)

[Young Drachma]

Simply put, that game ending was a bug. Arizona State was ranked #2 in the country, UNO were a fluke but didn't die. It makes no sense they kicked a field goal in their last possession, I don't have pbp game logs so it's unclear how much time was left but obviously the game logic doesn't have any idea the stakes, what game it is, or even what to do. I have 2 minute drill logic built into this, I've expanded the coaching system pretty considerably.

But the end of game mechanics including hail marys and stuff were not things I've ever built into this so far, and I'd honestly never really paid much mind to an individual box score and game log enough to notice something that silly.

So I'll get that patched up. Even if UNO would've won and kept their Cinderella story, game ending FG made no sense.

What's been fun and overwhelming about building this is when I test stuff like this out and i'm trying to optimize for volume, and frankly, seeing how real the sport feels to me. I still haven't nailed down the "house style" for writing about it, LLMs have a tough time translating it from box scores and it wasn't to narrate too much, but seeing it do that and reading other stuff online has given me my own sort of interpretive voice about it myself which kind of helps and I feel like I'm starting to get there.

I've also updated the game engine to essentially see "can I actually win this game? What do I need to do?" versus how it was before where it was just opting for field goals because they're easier when other stuff isn't working.

There was another instance with Arizona State kicked a field goal on 3rd down -- why -- and the engine let it happen, it was 3rd & 17 but still that makes no sense with more clock, more downs and sure you need points but WTF. So fixing that logic, too.

[Young Drachma]

One of the things I've been trying to do with this besides simulate seasons is find all of the bugs. It's just a simulator, we're not trying to make a game here that I can ever play, that'd be too much work and even just building a simulator over the past few weeks off and on has been a lot of work.

It's been fun slash interesting to find various bugs inside the engine and things that defy football logic like clock management. At the core, having a game that's essentially football with some cooked up rules makes it easier to have a mental model for what I'm actually trying to do here, versus changing the physics of the sport or something else.

The other thing I've been contending with is what it's supposed to sound like when you write about it. Because i'm simulating, I'm looking at box scores and trying to reverse engineer from the drive summaries and stats what's actually happened in the game. Some of that has led to lots of changes in the engine including fixing bugs, creating more distance between elite teams, good teams, average teams and legitimately bad ones, because there was too much closeness in talent that was creating havoc on playoff matchups.

Increasingly, it's started to feel more real though. I can look at results, see stats and understand what's going on. I think contextuaizing it in the ways that I view other sports has been the latest thing for me to try to do, I don't know why I care, but it's been sort of fun to take this from concept to feeling more like it's closer to tangible.

I might write a cleaner dynasty that's just a look at a season so I can talk about the sport, its mechanics and kind of what happens and why it's different than what football looks like even though it rhymes.

[Young Drachma]

So many years ago, I had my own computer college football ranking. At the peak there were 100s of them (our original patron saint of PC football Jim Gindin had one too) and the basic premise is that college football has a lot of teams, not very many game and so how can you pick a champion with those conditions?

How do you know who the best team actually is?

Thankfully, there have been lots of math nerds over the years -- many who don't even care about football -- who were working with the data sets to try to "solve" this problem and Kenneth Massey had an online comparison of all of these rankings to essentially figure out who was the consensus of these models top team.

Realizing college viperball is just a different flavor of college football, I got the idea that there was probably enough data in the engine to use some of these same formulas and models for viperball too.

What we get ultimately is something even more interesting, because Viperball has a different set of asymmetric problems that make it worth trying to understand "is this team really lucky or are they good?" along with all of the other schedule/win-quality related problems that football has.

Tennis probably more fair than Viperball is, even if you can win a match without scoring more points than the other person. Viperball mimics some of this mechanic with Delta Yards at kickoff for trailing teams. Knowing the sport isn't real hasn't made the results feel less that way, I'm playtesting and debugging like you would for a normal game/sport but I'm more interesting in nailing the mechanics rather than "winning" or playing which is kind of my sweet spot for a sim.

The sim is python based, the GUI is an presentation layer are actually different but I always built it with the FastAPI knowing it could be extended at some point.

[Young Drachma]

So many years ago, I had my own computer college football ranking. At the peak there were 100s of them (our original patron saint of PC football Jim Gindin had one too) and the basic premise is that college football has a lot of teams, not very many game and so how can you pick a champion with those conditions?

How do you know who the best team actually is?

Thankfully, there have been lots of math nerds over the years -- many who don't even care about football -- who were working with the data sets to try to "solve" this problem and Kenneth Massey had an online comparison of all of these rankings to essentially figure out who was the consensus of these models top team.

Realizing college viperball is just a different flavor of college football, I got the idea that there was probably enough data in the engine to use some of these same formulas and models for viperball too.

What we get ultimately is something even more interesting, because Viperball has a different set of asymmetric problems that make it worth trying to understand "is this team really lucky or are they good?" along with all of the other schedule/win-quality related problems that football has.

Tennis probably more fair than Viperball is, even if you can win a match without scoring more points than the other person. Viperball mimics some of this mechanic with Delta Yards at kickoff for trailing teams. Knowing the sport isn't real hasn't made the results feel less that way, I'm playtesting and debugging like you would for a normal game/sport but I'm more interesting in nailing the mechanics rather than "winning" or playing which is kind of my sweet spot for a sim.

The sim is python based, the GUI is an presentation layer are actually different but I always built it with the FastAPI knowing it could be extended at some point.