Comparing Dyson's d6 Thief and LotFP's Specialist

TL;DR: give LotFP Specialists 8 skill points to invest, not 4.

The celebrated Dyson Logos has a post on thieves that's nearly a decade old now. In the post he takes the B/X Thief skill table, which was given as percentile rolls, and converts them to their d6 (X-in-6 chance) equivalents.

(There was very little reason to present Thief skills as percentile rolls in the first place. All of them were in 5% increments, bar Climb Sheer Surfaces which ranges from 87% to 99%, and Hear Noise, which was given as an X-in-6 chance. Why, TSR?)

Lamentations of the Flame Princess, famously, also uses the d6 for skills. They give all characters a 1-in-6 base chance at succeeding, and allow Specialists (Thieves) to invest 4 skill points in any of them, and a further 2 points on each level up.

Converting Dyson to LotFP

So what would Dyson's numbers look like in LotFP? In the table below I've taken his numbers, changed the skill names to their LotFP equivalents, and removed Hear Noise. I've also merged Pick Locks with Find or Remove Traps, and Move Silently with Hide in Shadows, always choosing the higher odds of the two merged skills. Sneak Attack is a skill in LotFP, but originally was a class feature fixed at 2x, so I've set that at 2. The new LotFP skills not based on Thief features are omitted. Finally, we tally up the total "skill points" in this converted-Dyson system, and compare that number to the skill points from LotFP's Specialist class table.
 
Dyson conversion vs. LotFP

Level	Tinker	Sleight of Hand	Stealth	Climb	Sneak Attack	Total skill points	LotFP Specialist skill points
1	1	1	1	5	2	5	4
2	1	1	1	5	2	5	6
3	1	2	2	5	2	7	8
4	2	2	2	5	2	8	10
5	2	2	2	5	2	8	12
6	3	3	3	5	2	11	14
7	3	3	3	5	2	11	16
8	4	4	4	5	2	14	18
9	4	4	4	5	2	14	20
10	5	5	5	6	2	18	22
11	5	5	5	6	2	18	24
12	6	6	6	6	2	21	26
13	6	7	6	6	2	22	28
14	6	7	6	6	2	22	30

Dyson agrees with most people that the starting values are much too low. In the post he also presents a conversion to 2d6, and a houseruled version of it to boost the skills. (The houserule even includes choosing one "favoured" skill - could this post be present in LotFP's DNA?) Here, I won't be looking at the 2d6 tables. Instead, let's find out how the d6/LotFP skill conversion of the (famously stingy) B/X table compares to the LotFP ruleset. Starting from a base value of 1, the above table has 4 points invested into Climb, and 1 into Sneak Attack. That makes for a total of 5 skill points - more than the 4 that LotFP grants to first level Specialists.

The two rightmost columns of the table show that the B/X Thief is overtaken by the Specialist in skill points at 2nd level. To be fair to LotFP, though, we did chose the higher of two skills when merging them, instead of taking their average. It may also be argued that the LotFP Specialist's bump to a d6 hit die, and the ability to choose which skills to improve, offsets this loss in skill points.

A Little Boost

If, then, we agree with Dyson and others that these numbers need a boost, particularly at low levels, what would it look like under LotFP rules? Since all classes have skills at 1-in-6 in LotFP, the smallest reasonable buff we can give the Thief/Specialist is to start all Thief skills at 2-in-6 instead of 1.

Boosted Dyson conversion vs. LotFP

Level	Tinker	Sleight of Hand	Stealth	Climb	Sneak Attack	Total skill points	LotFP Specialist skill points
1	2	2	2	5	2	8	4
2	2	2	2	5	2	8	6
3	2	3	3	5	2	10	8
4	3	3	3	5	2	11	10
5	3	3	3	5	2	11	12
6	4	4	4	5	2	14	14
7	4	4	4	5	2	14	16
8	5	5	5	5	2	17	18
9	5	5	5	5	2	17	20
10	6	6	6	6	2	21	22
11	6	6	6	6	2	21	24
12	6	7	6	6	2	22	26
13	6	8	6	6	2	23	28
14	6	8	6	6	2	23	30

Now, the score at 1st level is a whopping 8 points to LotFP's 4. The Specialist only overtakes the Thief at 5th level. However, in my opinion, the levels below 5th are far more important to the game than the ones above it, so the LotFP skill point allocation seems lacking.
 
My conclusion is to keep LotFP Specialists otherwise as they are, but try giving them 8 skill points at 1st level instead of 4. (I won't limit the amount of skill points invested into any one skill, at least to start with. If there turns out to be a problem allowing one to max out one of their skills at 1st level, then we'll deal with that as it arises.)

Or, for a super simple rule: start Thieves at 2-in-6 for all skills (except Climb at 5-in-6), and increase this on levels divisible by three: to 3-in-6 at 3rd, 4-in-6 at 6th, and 5-in-6 at 9th level. Then, if playing beyond name level, bump everything to 6-in-6 at 10th.

Below is a comparison of the skill point equivalents of our various conversions with LotFP rules-as-written, as well as LotFP with a buff of 2 or 4 additional skill points at 1st level.

Comparison of skill point equivalent totals

Level	Dyson	Boosted Dyson	Dyson Boosted Smoothed	Super Simple	LotFP	LotFP+2	LotFP+4
1	5	8	8	8	4	6	8
2	5	8	9	8	6	8	10
3	7	10	10	11	8	10	12
4	8	11	11	11	10	12	14
5	8	11	12	11	12	14	16
6	11	14	14	14	14	16	18
7	11	14	16	14	16	18	20
8	14	17	17	14	18	20	22
9	14	17	18	17	20	22	24
10	18	21	21	21	22	24	26
11	18	21	21	21	24	26	28
12	21	22	22	21	26	28	30
13	22	23	23	21	28	30	32
14	22	23	23	21	30	32	34

I suggest you try out giving this higher number of skill points to Specialists/Thieves - whether your taste is to allow them to invest the points freely, or make a rigid progression resembling the B/X originals with Climb being far ahead of the others. See if that makes Thieves feel too overpowered and breaks the game - I doubt it will.

One caveat: starting Specialists with 8 points does mean that Thief skills are maxed out at 10th level, and only Sleight of Hand (which is modified by the target's level, and thus a base value higher than 6 still makes sense) increases beyond that. But that's only if you invest in Thief skills and nothing else; LotFP has other skills such as Architecture and Bushcraft to soak those points in. And honestly, I don't plan on playing beyond 10th level anyway.

Simple OSR Thief skills on a d20

Observation 1: Most of the Thief skills in B/X D&D improve in increments of 5%, which makes using a percentile die rather pointless. You can take the percentile (and x-in-6) Thief skills table, convert all the values onto a d20 target number, and it looks like this (TN rounded up):

Level	Climb Sheer Surfaces	Find/Remove Traps	Hear Noise	Hide In Shadows	Move Silently	Pick Locks	Pick Pockets
1	3	18	14	18	16	17	16
2	3	17	14	17	15	16	15
3	3	16	10	16	14	15	14
4	2	15	10	15	13	14	13
5	2	14	10	14	12	13	12
6	2	12	10	13	11	11	11
7	2	10	7	11	9	9	9
8	2	8	7	9	7	7	7
9	1	6	7	7	5	5	5
10	1	4	7	5	3	3	3
11	1	2	4	3	1	1	1
12	1	1	4	2	1	1	0
13	1	1	4	1	1	1	0
14	1	1	4	1	1	1	0

Observation 2: From this table it is easy to see that most of the skills - Find/Remove Traps, Hide in Shadows, Move Silently, Pick Locks and Pick Pockets - are always within a band of 2 points of each other. Generally, they start at 16-18, improve 1 point per level, before accelerating to a rate of 2 points per level starting at 5th or 6th level. These could very easily be collapsed into one chance without losing much precision - let’s say into the middle chance, Pick Locks, a 17.

One of the remaining two skills, Hear Noise is one that many tables including mine (and some retroclone rulesets) ignore, simply letting players succeed to listen if they spend a bit of time, so it can be ignored here. Climb Sheer Walls is a bit of a tricky one. It’s the only one that starts at a high chance of success, 87%, and improves by one percentage point per level, so it ill fits non-percentile models. However, there is precedent in the OSR of merging Climb into a more “unified” skill system, in LotFP, where Climb improves in steps of the same size as other skills. If Specialists put all their points at 1st level towards Climb, they end up with chances quite similar to the ones in B/X - good chances in climbing (5/6), terrible chances of doing anything else (1/6).

Hacking it

Let’s assume that in our modern mindset we want a resolution mechanic that can be applied not just to Thieves, but all classes, while still retaining Thief superiority. Naturally, we achieve that by treating all other characters as zeroth level Thieves. So, ignoring Climb for a moment, we choose a 17 as our target number at 1st level, increasing by one each level. In other words, to succeed at a Thief skill, you must roll d20 + Thief level for a result of 18 or higher.

Hell, why not call it a nice round easy-to-remember 15, and give Thieves a little boost? The B/X Thief is commonly seen as too inefficient, and in fact AD&D boosts many of the thieving chances by 10%. Now, doing this does improve all characters’ odds from the default 16.6% of LotFP to 30%. But, especially these days, most OSR play takes place in the low levels. I’m okay with letting characters do a bit more. (For comparison, the Mutants & Magic proposal linked above lets 1st-level Thieves succeed on most things on a 12-14+.) Incidentally, choosing 15 as the target number, the new chances start to fall behind the B/X Thief's at 8th-9th level.

"Most LotFP campaigns feature lower level characters, and modern campaigns do not last long enough to build characters to higher levels. ... providing characters more possibilities at lower level and decreasing the importance of gaining levels is important."

- James Raggi, "Playtest Notes" in Eldritch Cock

Now that we’ve moved all skills to this mechanic, the obvious next step is other “common activities” with X-in-6 odds like Foraging, which LotFP turns into the Bushcraft skill, and Open Doors, which is not a skill but an X-in-6 chance modified by Strength. Open Doors should certainly be modified by Strength in our new system too. One benefit of moving to the d20 is that if you want ability scores to affect the odds of skills, you can, simply by adding the ability modifier to the roll. Wisdom would be used for Bushcraft, Dexterity for Stealth, Intelligence for Tinker, and so on. The modifiers have a bit of an impact but do not overflow the die size. Also, we are now rolling high (like all other rolls), instead of rolling low on a d6 or percentile, if you care about that sort of unification/elegance nonsense. (I don't think the d6 is bad at all, personally, in fact I really like how chunky the increments feel.)

Finally, let’s get suspiciously modern and use the term “proficiency” for adding Thief levels to the roll. Thieves shall have proficiency in Stealth, Climb, and Tinker. If we take a leaf from LotFP’s increasing demihuman skills, then Dwarves would have proficiency with Architecture, Elves with Search, and Halflings with Bushcraft (and also have a +10 to hide in the wilderness). I would also probably give Fighters proficiency in Open Doors, because why not? They’re likely to be the most physical type.

A summary of the skill system

All activities and skills (Climbing Sheer Walls, Finding Traps, Removing Traps, Hearing Noise, Hiding in Shadows, Moving Silently, Opening Locks, Picking Pockets, Forcing Open Doors) are rolled on a d20. The appropriate ability modifier is added, as well as Thief level if the activity is thiefy. A result of 15+ is a success.

Skills/activities:

• Architecture: INT, +Dwarf level
• Bushcraft: WIS, +Halfling level
• Climb: STR, +Thief level
• Languages, Medicine: INT
• Open Doors: STR, +Fighter level
• Search: INT, +Elf level
• Stealth, Sleight of Hand: DEX, +Thief level (these skills can be rolled into one)
• Tinker: INT, +Thief level

(Sneak attack is no longer a skill, but a simple 2x multiplier, possibly increasing with Thief level.)

Now that all Thief activities progress the same, you can combine those that use the same ability, i.e. Stealth and Sleight of Hand. If you see Tinker (Pick/Remove Traps) as an activity requiring more dexterity than intelligence, then that can be merged in too for a general "Thievery" skill. As an extension of the system, you could have character backgrounds that give you background skills granting a +2 bonus to certain activities.

Pros & Cons

So what’s the score? We have made abilities slightly more important, which depending on your taste may be a good thing or a bad thing (but actually less important compared to roll-under ability checks). We now have a Thief and not a Specialist - we’ve removed “builds”, option paralysis, customization - which again you may see as either good or bad. We've made Thieves decent at everything, but far worse at climbing (I dunno, give them a +10 to climb or something if you want). As a positive, we’ve reduced the space taken up by skills on the character sheet, since you no longer need to list the individual skill levels, only whether you’re proficient in them - adding your level to the roll is easy enough. And there's no need to print a Thief skills table in the rulebook anymore.

However, I see some serious risks in going in this direction. Those risks are psychological shifts that may happen in dungeon masters when moving from a “X in Y” chance to a target number which has treacherous resemblance to the modern “d20 + proficiency vs. difficulty class” mechanic of 5e. The danger is that DMs will start to see the fixed target number of 15 as a "default DC" that they will begin to carelessly modify. They may be tempted to do this where they would not have been in a d6 skill system. LotFP's rules only mention modification of the chances in two places: opening giant stone doors, and attempting to forage in different terrains such as the desert. Modifying the target numbers too often may give rise to a treadmill where characters get more competent but DMs constantly raise the obstacles in response. Because... for some reason a locked chest in a high-level dungeon would have a higher DC to open? I am generally against that, and it should be used very sparingly, as LotFP does. Just let players know their chances of success, and let characters feel competent! This psychological risk of treadmills could perhaps be mitigated by pre-applying the modifiers and listing the naked die result required on the character sheet, as saving throws are handled, though that would take up a bit more space on the sheet and require updating every level.

The other risk is that it will tempt DMs to call for knowledge rolls, and perception rolls. Those are harmful too, but the explanation why is a subject for another blog.

Finally, I'd like to say that I'm not under the impression any of this is an original idea. The Thief seems to be one of those topics of eternal discussion. In fact I'm almost certain there are already several blogposts with same idea, but Google failed to find me any so I'm barfing my notes on here.