- Thread starter
- #26
Meant to add this at the beginning. Here is the carry dispersion chart after removing all of the fat shots. I have no idea why, but the Ai150s blow the scale up.
Dan.The.King's AI fitting help
Very cool - I love how intuitive these results are (even if they seem “off” in any way). Def agree that refining the prompt a bit could give you different results, but I actually agree with some of its takeaways on stability & dispersion being better in a more stout shaft.I told it tight dispersion was most important to me.
Best setup for tight dispersion (from what you tested)
#1 — Callaway Apex AI150 + MMT 125 Stiff
This is your winner for dispersion. Full stop.
Why this combo works for you:
- 125g shaft stabilizes face delivery at your ~83 mph 7i speed
- MMT profile is smooth but doesn’t kick unpredictably
- AI150 head:
- Compact enough to reduce gear-effect misses
- Forgiving enough to protect heel/toe strikes
- Typically produces:
- Slightly lower launch
- Controlled spin
- Narrow left/right window
If dispersion is king, this is the safest gamer from the fitting.
Runner-up (very close)
Srixon ZX5 + Nippon Modus 120 Stiff
This is the “metronome” setup.
Why it’s so good:
- Modus 120 is one of the tightest-dispersion steel shafts ever made
- ZX5 head keeps ball speed up on mishits without spraying
- Excellent front-to-back consistency, not just left/right
What likely hurt dispersion in the others
MMT 85 (Titleist T150)
- Too light → timing variability
- Increases closure rate → left/right misses
- Often looks great on single swings, bad over 10 balls
MMT 105 (Mizuno 225)
- Not terrible, but:
- More face rotation than 120–125g
- Slightly wider pattern over a full session
If you want to maximize dispersion control, here are two better-than-stock tweaks I’d strongly consider:
How to tighten dispersion even more (important)
If staying graphite:
Shaft tweaks (biggest gain)
If open to steel (often tighter):
- MMT 125 Stiff (best graphite choice you hit)
- Modus 120 Stiff
- Dynamic Gold Mid 120
- Project X 6.0 (only if you liked a firmer feel)
With your swing profile:
This alone can shrink your left-right window by 20–30%.
- 1° flat is very often the dispersion unlock
- Especially if your miss pattern includes:
- Pulls / pull-fades
- Left misses when you “step on one”
My “tour van” recommendation
If I were building this for you with dispersion as the #1 KPI:
One last check (quick but important)
When you missed during the fitting, which happened more often?
That answer lets me lock in lie + shaft profile with zero guesswork.
Left
Right
Mostly short/long
Again, dispersion was better with the MMT 105, so I'm not sure where it's going. Also, my main miss is a push right, which I told it after this, so I know going flat is not a good idea.
- Thread starter
- #28
I agree. I'm not going to throw everything out, but I do need to refine what I'm asking. Instead of just looking at dispersion I need it to refine the best choice based on more categories.
Maybe I'll ask it to project what the results would be if they were all coming in at a similar attack angle since I was definitely getting shallower as the session went on.
Did you try the AI 200s? If you liked the zx5s is another that may deliver good results
- Thread starter
- #30
I tried them in an impromptu fitting a few months ago. I thought the 200s would be the best for me in that line, but the fitter said I was more consistent with the 150s even if I wasn't getting all of the performance of them. I don't know why we didn't try them, but after the zx5s every club was kind of that step up.
I've done a lot with AI over the past year or so. A few things I've realized without its own deeply integrated and personalized knowledge base of data and a specific coaching or fitting philosophy it's basically just spitting back all of the random knowledge found on the internet in reviews and posts from forums like this one.
I think sometimes it gets lost in translation that all "out of the box" AI does is predict the most probable outcome of words based on information present on the internet. It's not inherently smart or an "expert" in anything.
It can start to become specialized if you create a specific model trained on very particular parameters but then all you are building is a confirmation biased machine that responds inside of the little bubble of preferences and biases you baked into it with the material you trained it on.
It's fun to mess around with. I've found a lot of utility in the model I've trained when it comes to helping me analyze the data I feed it and come up with ways to make that data better, but I'd never ask it to do a fitting for me. There are way to many variables it just can't see and I'd much rather have a highly trained human who is in the room with me observing things like tempo, club delivery, ball flight, and all the other 'intangibles'.
It's good for data collection and analysis. Those things are what AI excels at. Not this.
I think sometimes it gets lost in translation that all "out of the box" AI does is predict the most probable outcome of words based on information present on the internet. It's not inherently smart or an "expert" in anything.
It can start to become specialized if you create a specific model trained on very particular parameters but then all you are building is a confirmation biased machine that responds inside of the little bubble of preferences and biases you baked into it with the material you trained it on.
It's fun to mess around with. I've found a lot of utility in the model I've trained when it comes to helping me analyze the data I feed it and come up with ways to make that data better, but I'd never ask it to do a fitting for me. There are way to many variables it just can't see and I'd much rather have a highly trained human who is in the room with me observing things like tempo, club delivery, ball flight, and all the other 'intangibles'.
It's good for data collection and analysis. Those things are what AI excels at. Not this.
Have been following these threads and also trying my own when I have time or in sim rounds playing with 2 drivers just trying to optimize head setup/hosel settings with each. ChatGPT for example is what I'm using and it seems to get stuck or often run off with something. It is only going to do what you ask of it and the better parameters you give it to follow the better it can work for you.
Ive atleast found that if i tell it how an adjustment felt to me and how the session was swing wise as if I was having an off day or getting fatigued towards the end. It will look at the results differently and weigh it differently so it starts to recognize that.
Ive atleast found that if i tell it how an adjustment felt to me and how the session was swing wise as if I was having an off day or getting fatigued towards the end. It will look at the results differently and weigh it differently so it starts to recognize that.
AI hallucinations and drift are real thing when you go down threads. Often time you are best leaving the one question and starting over again but with different prompts and data. If you keep going it gets worse and worse.
- Thread starter
- #34
I don't disagree with anything you said. I'm not trying to use this as a "this is the club head and shaft I should blind order" but more as a fun little exercise to see what it spits out and to see if there are options I could try in an actual fitting. I also don't have the amount of data to input that you and others do.
Good deal. Yeah like I said it's a fun to mess around with. But especially when you get into something like shafts. The AI has no idea what your Tempo is and what matches up well with your swing. All it's doing is giving guesstimates based on your swing speed. The same guesstimates that all of us online usually get wrong until we go in somewhere and get fit.
There's a whole thread here on thp about people purposefully going to lighter iron shafts and seeing great results. I'm not saying that could be you, because it's definitely not me, but the AI is not going to know that for either of us.