Using AI to Generate Player Scorecards (1 Viewer)

[thePokerClub]

How I Use AI to Generate Player Scorecards
After each event, I use Claude to generate personalized PDF scorecards for each player in my club. The whole process takes about 10 minutes for 15+ players.

Wanted to share the workflow in case anyone else wants to try this.

---

What's in each scorecard:
Current season standings and points breakdown (wins, cashes, bounties, participation)

• Personal performance trends across the season
• Event-by-event results timeline
• Head-to-head knockout records
• Projected path to Championship qualification
• "What it would take" scenarios to reach the next LeaderBoard position

---

Sample output (my own scorecard):

thePokerClub Race-to-100 Performance Scorecard - Chris Sawyer

---

The workflow:

1. After each event, I export my BlindsUp data and update my master tracking spreadsheet with results
2. Export the relevant data as CSV
3. Feed it to Claude with a prompt that specifies the scorecard format I want
4. Claude generates individual summaries for each player
5. I drop each into a simple PDF template and send privately

The first time took some back-and-forth to get the prompt right. Now it's basically copy-paste.

---

Why I do this:
My players love it. Gives them something to look at between events, fuels the group chat trash talk, and keeps the season feeling competitive even for guys in the middle of the pack.

It's also surfaced insights I wouldn't have noticed manually - like which players are "due" based on their historical patterns, or who has a brutal head-to-head record against someone else.

I am actually thinking about a future incentive that multiplies bounty points if someone knocks out their 'Nemesis'. So, this process has definitely spurred more innovative thinking which is always great.

---

Happy to share:

• The prompt I use
• The PDF template
• Tips on what data to track to make this work

Just drop a reply or DM me.

-Chris

 

[HiveKueen]

Do you think this fosters wanting to learn more about playing the game? I'm always seeking ideas about how to nurture a learning/improvement approach to the game. Some of my players have it but some don't.

For example, I like to put out a "How many outs?" scenario for players to puzzle over and discuss. Some players really get into it and others don't even look at it. I'm pretty sure the word "out" was new to some of the players the first time i did this.

Recently i put out a small library of poker books for people to borrow. I have noticed one player look them over.

I think half my players would really like this score card and half would probably be alienated by it. Do you have any players who don't want it?

Cheers,
HiveKueen

 

[Poker Zombie]

I give up.

What's an LBP?

 

[raynmanas]

I give up.

What's an LBP?

my guess would be Leader Board Points.

 

[thePokerClub]

I think half my players would really like this score card and half would probably be alienated by it. Do you have any players who don't want it?

I don't have any players that have said they don't want it.

The majority say, 'Thanks, this is cool!'. But, does that mean they understand the control chart at the top? Maybe/maybe not. In my mind that is potentially the most confusing piece. What I was going for was a Schewart control chart that considers overall performance of the group and then sets an acceptable performance range to be in (or above) consistently in order to be part of the Top 8 (number of players that will qualify for our Championship). This is the gray box horizontally across the time series.

The rest is fairly straightforward.

I get most compliments or 'I really like's' about the Nemesis callout, and the impact of Rebuys and Add-ons on overall group performance. I like disproving that just because someone Rebuys multiple times it improves their win-rate...my general philosophy is that this is directional incorrect thinking. And the data for this group this Season proved that out. The recommendation that everyone always do the Add-on because that was more statistically significant as far as LeaderBoard Point accumulation was interesting.

Not many comments about the last item which is the attempt at a 'roadmap to Top 8'.

This was my first pass at using AI to generate what I had envisioned as a scorecard. I will keep tuning it up but for now, this is what I will roll with.

As it always is with data - it is an iterative evolution and understanding requires consistent messaging and time.

 

[thePokerClub]

my guess would be Leader Board Points.

Yup, LeaderBoard Points.

 

[HiveKueen]

I like disproving that just because someone Rebuys multiple times it improves their win-rate...

Interesting, my stats show clearly that rebuying when possible after a bust greatly improves chances of getting into the season championship. We don't do an add-on as this would make the game last too long for a weeknight.

 

[thePokerClub]

Interesting, my stats show clearly that rebuying when possible after a bust greatly improves chances of getting into the season championship. We don't do an add-on as this would make the game last too long for a weeknight.

To clarify further, our data showed that multiple rebuys did not help players performance, but a single well-timed rebuy did.

In other words instances of a single rebuy for the night in a certain Round aligned with better performance. Probably because of the number of BBs someone came in at off of that.

Rapid fire risk taking for the purpose of rebuying was not helpful to performance.

 

[Poker Zombie]

I don't use AI, but I also build a report. I'm into the 3rd version now, as I think seeing the same thing every time would be a waste of paper.

 

[Jonesey07]

I don't use AI, but I also build a report. I'm into the 3rd version now, as I think seeing the same thing every time would be a waste of paper.

View attachment 1639665
View attachment 1639674

Now me!!

 

[thePokerClub]

I don't use AI, but I also build a report. I'm into the 3rd version now, as I think seeing the same thing every time would be a waste of paper.

This is impressive historical depth - 126 events over 16 years is a dataset most clubs would kill for. Do you do this on all players? Likely would only happen if you are the host I guess.

The "Percent of Field Defeated" chart with the 8-game moving average is the piece that jumps out to me. That's actual trend tracking.

Curious, have you done anything with the positional data (who you act before/after)? Seems like there could be insight buried there - like whether your cash rate changes based on who's on your left.

Appreciate you sharing this. Cool to see how other clubs are thinking about player data.

 

[Poker Zombie]

I haven't "done" anything with positional data, pre se. I once heard someone say "You always sit on my left", so I decided to track it to see if there was any truth to the matter.

We do "weight" table assignments. 1/3 the weight is past performances over the last 2 seasons (and the current one). 1/3 the weight is the OtM1P (the percent of the time a player both finished out of the money and the eventual winner started at their table). 1/3 the weight is purely random. Players that do poorly start at an outside table, while winners tend to start at the main table. However, actual seat placement is true random - though the best finisher from the previous event starts at seat #1 (unless a player is physically disabled and needs easy access in which case seat 1 is rotated around the table so the random seating is preserved). Because of table weighting, you are more likely to start at a table with someone of a roughly equal skill level, and those players may sit to each other's left or right more frequently.

 

[Poker Zombie]

Now me!!